Following a recent report in the Metro that BT is carrying out research into continuous authentication software, we look at some of the pros and cons and the issues around employees potentially being filmed all day at their desks … under the guise of cyber-security.

Why Use Continuous Authentication Technology? 

Businesses use continuous authentication technology to enhance security, i.e. to add an extra layer of protection. As the name suggests, this type of software continuously verifies users throughout their session, rather than relying solely on traditional one-time authentication methods like passwords or PINs. This approach is designed to mitigate risks such as session hijacking, whereby unauthorised users gain access after the initial login, or insider threats where someone might misuse another’s logged-in session. Continuous authentication essentially helps detect abnormal behavior in real-time, flagging up potential breaches or fraud by monitoring unique patterns such as typing style, mouse movements, or facial recognition. By integrating this technology, businesses may hope to reduce security vulnerabilities, safeguard sensitive data, and improve compliance with industry regulations, all while maintaining a seamless user experience, i.e. it’s happening automatically in the background.

BT Trialling Continuous Authentication Technology 

BT is reported to be trialling BehavioSec’s behavioral biometrics technology at its Adastral Park science campus near Ipswich. This software is used for continuous authentication, where it monitors users’ unique behavior patterns, such as how they type, move the mouse, or interact with their devices, to confirm their identity. However, in the case of BehavioSec’s technology, it doesn’t usually require the use of a camera, i.e. the user doesn’t need to filmed by a webcam all day. Instead, it can rely on analysis of a user’s behaviour patterns by looking at factors such as keystroke dynamics, mouse movements, touchscreen gestures, and device interaction patterns (e.g. how the user holds their phone, scrolls through pages, or interacts with specific applications). In the recent Metro story however, the reporter witnessed a demonstration of the system that did use facial recognition and required continuous filming of the user with a webcam/front-facing camera to detect whether the user’s face was consistent with expected dimensions.

BT is exploring this technology as part of its broader efforts to improve cybersecurity, particularly in response to the growing threat of cyberattacks and data breaches. The trials of BehavioSec’s behavioral biometrics technology are part of BT’s research into how it can use innovative technology to better protect digital assets and infrastructure, especially in enterprise and government contexts. For example, back in 2022, BT said it would be taking security to a new level so that even if an attacker obtained a device, any ongoing work session would end, locking the device, because their biometrics wouldn’t match that of the device user’s known biometrics.

Systems Using Cameras? 

There are, however, many such continuous authentication systems now available which require a camera being trained on a user’s face. A few prominent examples include:

– FaceTec’s ZoOm. This is a 3D facial recognition solution that uses the front-facing camera of devices (it can use a webcam) to authenticate users, e.g. by carrying out “Liveness Checks, Face Matches & Photo ID Scans”. It’s often used in applications requiring high security, such as financial services or identity verification systems, and biometric security for remote digital identity.

– FacePhi. This (Spanish) biometric solution for facial recognition is widely used in the banking, healthcare, and fintech sectors for secure access to mobile banking apps and fraud prevention. The software uses a camera to identify users and offers continuous authentication by tracking facial features during interactions.

– IDEMIA’s VisionPass. This system combines 3D facial recognition with AI and uses cameras to recognise faces and continuously verify identities, even in challenging conditions like low light or with face masks. It’s generally deployed in secure facilities, airports, and government buildings for access control and ongoing authentication.

– Trueface. This AI-powered facial recognition technology integrates with existing security systems, such as cameras in corporate offices, to provide continuous authentication. Trueface can recognise and track users in real-time, improving access security and is used in corporate offices, airports, and law enforcement for continuous identification and authentication.

Other popular systems that use similar methods include Clearview AI, Neurotechnology’s Face Verification System, AnyVision, and ZKTeco’s FaceKiosk.

It’s also worth noting here that the “big tech” companies’ versions, such Apple’s Face ID, Google’s Face Unlock (Pixel Devices), and Microsoft Windows ‘Hello’ are also facial recognition-based authentication systems that are classed as continuous authentication technology. However, for the purposes of this overview, we’re focusing on the kinds of systems that businesses may use for their own employees.

Issues 

The usage of facial recognition (e.g. by law enforcement) has had its share of criticism in recent years. However, the thought of businesses using a camera to continuously film an employee, even if it may be for security purposes, such as continuous authentication, raises several serious issues and concerns. For example:

– An invasion of privacy. With constant surveillance, employees may feel that their privacy is being violated. Cameras can capture not only work-related activities but also personal moments, which may lead to discomfort and a sense of being micromanaged. Cameras might inadvertently record personal or sensitive information, such as confidential discussions, which could be accessed or potentially misused.

– The effect on employee trust and morale. Continuous filming can create an atmosphere of distrust between employees and employers. Workers may feel they are being monitored for reasons beyond security, leading to an atmosphere of fear, plus a decrease in morale and engagement (and ‘quiet quitting’).

– Psychological stress. Constant camera surveillance can lead to stress or anxiety among employees, affecting their overall well-being and productivity, which could obviously be counterproductive for the company.

– Data security and misuse. For example, video recordings of employees can contain sensitive biometric data, which, if compromised through a data breach, could have serious consequences. Biometric data is immutable, i.e. once stolen, it cannot be changed (like a password). There is a risk of video footage being misused, either by internal parties or external hackers. The footage could be exploited for purposes other than security, such as inappropriate monitoring of behavior or harassment.

– Ethical concerns. These could arise if employees are not fully aware of the extent and purpose of the surveillance, or if they feel coerced into accepting it as a condition of employment. Also, filming employees all day can be viewed as excessive (overreach), especially if less invasive alternatives exist. Monitoring behavior to this degree may cross ethical boundaries of acceptable workplace practices.

– Legal implications. Many regions have strict privacy laws (e.g. GDPR in Europe, CCPA in California) that require companies to obtain explicit consent for continuous surveillance and ensure the proportionality and necessity of such measures. Non-compliance could lead to legal consequences, fines, or lawsuits for a business. In some countries (or US states, for example) there are labour laws that protect employees from invasive workplace monitoring. Continuous surveillance may violate these protections if it is deemed too intrusive.

– The Potential for bias and discrimination. Among other things, this could include algorithmic bias. If the continuous authentication system relies on facial recognition, there is a risk of bias against certain groups, such as racial minorities or those with disabilities, due to known issues with facial recognition accuracy across diverse demographics. Also, employees may worry that the surveillance data could be used for purposes other than security, such as evaluating performance, which could lead to discrimination or unfair treatment.

– Technical reliability, e.g. false positives/negatives. Continuous authentication systems relying on cameras may fail, leading to false positives (unauthorised users being granted access) or false negatives (legitimate users being denied access). This can disrupt work and erode trust in the system.

While continuous authentication aims to enhance security, using cameras to film employees all day raises significant challenges. Companies need to carefully balance security needs with privacy rights, ethical considerations, and legal compliance to avoid potential negative consequences. For example, in 2020, H&M (the German multinational clothing retailer) was fined €35.3 million by the Hamburg Data Protection Authority in Germany for violating GDPR due to excessive and invasive surveillance of employees.

What Is ‘Emotional Analysis’ And Why Is It Causing Concern? 

Some continuous authentication software can now use ‘emotional analysis’. This refers to the use of AI to detect and interpret human emotions through cues like facial expressions, voice tones, or body language. Its purpose is to monitor and assess workers’ emotional states, such as stress, engagement, or satisfaction. It could help a business by providing insights into employee well-being and productivity, identifying signs of burnout or disengagement, and enabling management to respond proactively to improve workplace morale, increase efficiency, and enhance overall performance through better support and tailored interventions.

However, its usage also raises significant concerns around privacy, accuracy, and bias. The technology is often inaccurate, particularly across different demographics, leading to misinterpretation of emotions. Its use in workplaces for employee monitoring can create a sense of invasion and stress, eroding trust, and morale. There are also ethical and legal issues, with fears of misuse for micromanagement or even manipulation of behavior, making its widespread deployment highly controversial.

Susannah Copson, legal and policy officer with civil liberties and privacy campaigning organisation Big Brother Watch has described ‘emotion recognition technology’ as “pseudoscientific AI surveillance” and has called for it to be banned.

What Do Rights Organisations Say? 

Big Brother Watch is strongly opposed to the unchecked growth of workplace surveillance tools, calling them an invasion of privacy, harmful to employee well-being, and in need of stricter regulation to protect workers’ rights. Big Brother Watch recently held an event at the UK at the Labour Party conference to launch its report on workplace surveillance in the UK, highlighting its increasing use by bosses and their employers, and its negative effects on employees.

Big Brother Watch argues that workplace surveillance technologies, such as keystroke logging and AI-powered emotional analysis, invade employee privacy, erode trust, enable micromanagement, and harm mental health, potentially violating privacy laws like GDPR, while calling for stricter regulation to protect workers’ rights.

How Much Has Workplace Surveillance Increased?

A recent report by ExpressVPN, titled the “2023 State of Workplace Surveillance,” highlights a significant increase in workplace surveillance. Some key findings include:

– 78 per cent of employers are using some form of employee monitoring tools in 2023, up from 60 per cent before the COVID-19 pandemic.

– 57 per cent of employers implemented new surveillance tools specifically due to remote work conditions caused by the pandemic.

– 41 per cent of companies now use software to track keystrokes, screenshots, or record the activity of employees’ screens.

– 32 per cent of employers monitor employee emails and messages, while 25 per cent track employee location using GPS or IP data.

A Growing Market 

This surge in monitoring reflects the growing reliance on digital surveillance tools to manage remote workforces. Regarding the market for identity and access management (IAM) and cybersecurity solutions, Gartner reported in its “Market Guide for User Authentication” that continuous authentication is gaining traction due to increasing concerns about cybersecurity and the limitations of traditional login methods.

A MarketsandMarkets report has also noted that the global user authentication market, which includes continuous authentication solutions, is projected to grow from $13.9 billion in 2022 to $25.2 billion by 2027. A 2022 Verizon Data Breach Investigations Report also noted that 61 per cent of breaches involve stolen credentials and pushed companies to adopt continuous authentication as a preventive measure.

What Can Employees Do? 

If employees are concerned about continuous camera monitoring such as that used with some continuous verification systems, the (realistic) options they have are to:

– Review company policies to understand the purpose and limits of the surveillance.

– Raise concerns with HR or management to request less invasive alternatives, like fingerprint or password-based methods.

– Seek legal advice if monitoring violates privacy laws, or report it to a regulatory body like the ICO (in the UK).

– Consult with a union to negotiate privacy protections, if applicable.

– Document their issues for potential disputes and familarise themselves with their rights under local privacy and employment laws.

What Does This Mean For Your Business? 

The rise of continuous authentication software, particularly that using facial recognition and behavioural biometrics, highlights the tension between advancing cybersecurity and respecting employee privacy.

While the primary aim of these systems may be to offer ongoing, seamless security by monitoring users throughout their work sessions, the methods employed, such as continuous video surveillance or behavioural tracking, have raised significant ethical and privacy concerns. The promise of enhanced protection against cyberattacks, session hijacking, and insider threats is compelling, especially in industries where data security is paramount. However, the potential downsides of this technology can’t be ignored.

One of the key concerns is the invasion of privacy. Employees may feel uncomfortable or even violated if they know that cameras or other tracking mechanisms are monitoring their every move. The potential for these systems to inadvertently capture non-work-related activities, or even sensitive personal interactions, adds to the unease. Continuous surveillance risks creating an atmosphere of distrust between employers and employees, fostering a sense of being constantly watched, which could have a detrimental effect on morale. In extreme cases, this might lead to disengagement, lower productivity, or even a rise in ‘quiet quitting,’ as employees withdraw emotionally from their work due to feeling over-monitored.

Also, there are concerns about the psychological impact of constant surveillance. The knowledge that a camera or biometric system is perpetually tracking your behaviour can lead to stress, anxiety, and a feeling of being under perpetual scrutiny. This could, paradoxically, undermine the productivity gains that continuous authentication aims to protect. Employees working under these conditions might find it difficult to focus or perform optimally, especially if they perceive the surveillance as intrusive or excessive.

In addition to these privacy and security concerns, there are ethical and legal considerations. In many jurisdictions, privacy laws require companies to obtain explicit consent for such monitoring and ensure that the measures are proportionate and necessary. Failure to comply with these regulations could lead to hefty fines or legal action (as seen in the case of H&M’s €35.3 million fine in Germany).

There are also the issues of bias and discrimination. Facial recognition technologies have been shown to be less accurate across diverse demographic groups, potentially leading to unfair treatment of certain employees. If continuous authentication systems generate false positives or negatives due to these biases, it could create additional hurdles for employees from minority groups, further entrenching workplace inequalities. There is also the risk that the data gathered could be used for purposes beyond security, such as monitoring productivity or evaluating performance, which could lead to unfair assessments or discrimination.

Despite these challenges, it is clear why businesses are keen to explore continuous authentication technology. The ever-present threat of cyberattacks, data breaches, and insider threats has made it essential for organisations to find new ways to secure their digital assets. Continuous authentication offers a promising solution by providing ongoing verification without disrupting the user experience. However, businesses must tread carefully, ensuring that these systems are deployed in ways that respect employee privacy, comply with legal requirements, and avoid creating a toxic work environment.

As continuous authentication (seemingly inevitably) becomes more widespread, it will be crucial for businesses to engage in transparent communication with employees about how these systems work, why they are being implemented, and what safeguards are in place to protect their privacy. Offering alternative, less invasive methods, such as fingerprint recognition or password-based systems, may help alleviate some concerns. Ultimately, the successful adoption of continuous authentication will depend on striking the right balance between robust security measures and the protection of employee rights and well-being.

Following a recent report by cyber-security company Imperva about the rising costs to businesses of bot attacks and vulnerable APIs, we look at why it’s happening and what can be done.

Vulnerable APIs & Bot Attacks Costing Businesses $186 Billion 

Imperva’s report was based on Marsh McLennan Cyber Risk Intelligence Centre’s study of data from 161,000 cybersecurity incidents related to vulnerable APIs and bot attacks. The key findings were that businesses face an annual (estimated) economic burden of up to $186 billion due to vulnerable APIs and automated bot attacks. Also, the study found that these two security threats often work in tandem, are becoming increasingly prevalent, and pose significant risks to organisations worldwide.

APIs 

An API (Application Programming Interface) is a set of rules and protocols that allows different software applications to communicate with each other. Businesses adopt and use APIs because they enable seamless integration between apps and services, improving efficiency and automation. Mulesoft1 figures show that 99 per cent of organisations have already embraced APIs. An API can, for example, connect a company’s CRM system with its email marketing platform, thereby automatically syncing customer data. APIs also enhance customer experiences, like allowing users to log in via their Google or Facebook accounts. They help with scalability, such as a small business using cloud storage services via APIs to expand without building infrastructure. By using APIs for payments (like Stripe) or shipping (like FedEx), businesses can quickly innovate and offer services without developing them in-house. APIs also enable secure data sharing, such as a fintech company offering real-time stock market data through an API, while fostering partnerships, like travel booking sites combining flight, hotel, and rental services from different providers. This makes businesses more agile, efficient, and competitive in a connected world, thus highlighting positive outcomes of API adoption.

The financials of using them are illustrated by Mulesoft1 figures which suggest that many organistaions which are using APIs are reporting increased revenues, e.g. up to a 35 per cent increase, plus they are reporting reduced operational costs.

Why Are APIs Vulnerable? 

APIs are particularly vulnerable because they expose numerous endpoints, each acting as a potential entry point for attackers. As businesses increasingly adopt APIs to improve agility and efficiency, the number of these exposed endpoints has surged—on average, enterprises managed 613 API endpoints in 2023. This rapid expansion has created a larger attack surface, making APIs an attractive target for cybercriminals.

Also, with enterprise sites handling 1.5 billion API calls annually, the sheer volume makes the likelihood of encountering vulnerabilities greater.

What Kind of Vulnerabilities? 

The kind of business logic vulnerabilities in APIs include, for example, weak authentication, insufficient access controls, and improper data validation, all of which can allow attackers to exploit these APIs, leading to data breaches or system compromises.

What’s The Link Between Vulnerable APIs and Bot Attacks? 

Put simply, the link between vulnerable APIs and bot attacks is that:

– Greater API adoption (and a growing reliance upon them by organisations) has expanded the attack surface.

– Cybercriminals have realised that automated bots are a great and inexpensive way to attack the increasing number of vulnerable APIs, due to the scalability, speed, and efficiency of automated bots. Imperva, for example, highlights the fact that even low-skilled attackers can launch sophisticated bot attacks.

– Bots can quickly exploit multiple API endpoints – averaging 613 per enterprise in 2023 (Marsh McLennan) – making them ideal for large-scale attacks. Their low cost and 24/7 operation allow cybercriminals to probe for weak spots continuously, extracting sensitive data, executing fraudulent transactions, or launching disruptive denial-of-service attacks. Also, vulnerable APIs often lack strong security measures, thereby making them easy targets for bots, which can monetise stolen data or cause significant disruptions. As API adoption grows, bot attacks offer cyber-criminals a high-reward, low-effort method for exploiting these weaknesses, contributing to billions in annual financial losses.

This is why the Marsh McLennan Cyber Risk Intelligence Centre figures featured in the report show an 88 per cent rise in bot-related security incidents in 2022, followed by another 28 per cent increase in 2023. In essence, the more vulnerable APIs there are, the more bots are being used to attack them and as APIs become more integral to business, they become prime targets for bot attacks.

More Sophisticated 

One other key point highlighted in Imperva’s report is that the increasing sophistication of bad bots is a growing concern. For example, Imperva reports that over 60 per cent of bad bots detected today are classified as evasive, i.e. they use a mix of moderate and advanced techniques to carry out attacks. Worryingly, these bots can now mimic human behaviour, leveraging AI and machine learning to adapt and evolve over time. They can also delay requests and bypass common security measures like CAPTCHAs, making them harder to detect. This allows them to launch significant attacks with fewer requests, thereby reducing the typical “noise” associated with bot campaigns, making their actions stealthier and more effective.

The Financial Toll 

As mentioned at the beginning of this article, bot attacks on APIs are contributing significantly to financial tolls for organisations – up to $186 billion annually, with API-related breaches costing organisations up to $87 billion annually – an increase of $12 billion from 2021. Specifically, automated API abuse by bots now accounts for a massive $17.9 billion of these losses each year, thereby illustrating the immense economic impact of API vulnerabilities combined with bot-driven attacks.

Biggest Companies At Highest Risk 

Reseach appears to show that large enterprises (those with over $100 billion in revenue) face the greatest risk, with bot-related incidents making up as much as 14 per cent of all cyber incidents. Imperva’s report attributes the fact that they’re prime targets to their high visibility, extensive digital presence, and valuable assets.

Global Vulnerability 

Imperva’s report also highlights the global nature of API and bot attack threats, with countries like Brazil, France, Japan, and India now seeing high percentages of security incidents related to insecure APIs and bot activity. Although the proportion of such events in the United States is lower compared to these countries, the U.S. still accounts for 66 per cent of all reported incidents, highlighting its significant exposure to these growing threats.

What Does This Mean For Your Business? 

The financial and operational costs of API and bot attacks are escalating at an alarming rate. With global losses reaching as high as $186 billion annually, these threats are becoming a major concern for organisations of all sizes. The rapid adoption of APIs, while improving efficiency and agility, has also expanded the attack surface, making businesses more vulnerable. Automated bots, with their scalability and increasing sophistication, are exploiting these vulnerabilities at an unprecedented rate. Imperva’s report, featuring the findings of Marsh McLennan Cyber Risk Intelligence Centre’s study, appear to illustrate the severity of the situation. This situation appears to be worsening too as bots become more evasive, using advanced techniques like AI and machine learning to mimic human behaviour, evade detection, and carry out stealthy, highly effective attacks.

Larger enterprises, with extensive digital infrastructures, are particularly exposed, with bot-related incidents accounting for up to 14 per cent of all cyber incidents. These companies face significant financial risks due to their high-value assets and complex API ecosystems, making them prime targets for automated bot attacks. That said, smaller businesses are also frequently targeted due to potentially weaker security measures, meaning that businesses of all sizes should sit up and take notice.

It also appears that this threat is global, e.g. countries like Brazil, France, Japan, and India have experienced surges in API and bot-related incidents (although the U.S. remains the most affected).

As the digital landscape evolves, the overlap between API and bot vulnerabilities highlights the critical need for businesses and organisations of all kinds to adopt proactive, comprehensive security strategies. Businesses must tailor their defences to the specific risks associated with their size and complexity. For example, large enterprises managing hundreds of API endpoints need robust API security testing frameworks that regularly assess vulnerabilities, ensuring all endpoints are secure. This could include adopting authentication mechanisms like OAuth 2.0 or implementing rate limiting to restrict how many requests can be made to the API in a short period, which helps prevent bot-driven attacks.

Smaller businesses may want to focus on securing their APIs with proper encryption and multi-factor authentication to minimise exposure. They can deploy web application firewalls (WAFs) with bot management features, such as those provided by services like Cloudflare or Imperva, to detect and block malicious bot traffic before it reaches critical endpoints.

Both small and large businesses should adopt continuous monitoring for abnormal behaviour and invest in AI-powered security tools that detect patterns characteristic of bot activity. Also, penetration testing should be part of regular security audits to simulate attacks on API endpoints, exposing any weaknesses before they can be exploited by cybercriminals.

Munich-based Quantum Systems (a Small Unmanned Aerial Systems – ‘sUAS’ – company), has announced a successful test of AI-powered ‘drone swarm’ technology which could advance the role drones play in warfare.

What Is A ‘Drone Swarm’? 

In short, Drone swarm technology involves coordinating multiple drones to operate as a unified system and it can be used for tasks like military operations and surveillance, although it can be used for less sinister missions such as search and rescue and agricultural monitoring.

Test  

The announcement by Quantum Systems (working with Airbus) follows “intensive research” into whether it could “develop innovative solutions for the AI-supported autonomous control of swarms of drones” and “maximise the potential of artificial intelligence to coordinate mixed UAS swarms”. More specifically, research focused on the development of a Tactical UAS (Unmanned Aerial System), i.e. a system using small to medium-sized drones used for military operations (reconnaissance, surveillance, target acquisition, and communication support on the battlefield).  

Seven Successful Tests 

At a recent presentation at the Airbus Drone Centre in Manching (Germany), Quantum Systems announced that it had carried out seven successful tests of its AI-controlled UAS at the centre. Quantum Systems says its “new technologies enable large swarms of autonomous UAS to be effectively controlled by a small number of operators, even in highly dynamic and interference-prone environments.”

What’s So Different About The New Drone Swarm Technology? 

Drone swarm technology is already in use, particularly in military and defence applications in countries like the US, China, and Russia. Also, they’re used for civilian uses too, such as disaster response, search and rescue, agriculture, and entertainment (such as drone light shows).

However, following what’s been described as “a major breakthrough in autonomous swarm technology”, the drone swarm technology developed by Airbus Defence & Space, Quantum Systems, and Spleenlab (a software company) stands out for its use of advanced artificial intelligence (AI) and machine learning, particularly in autonomous coordination and decision-making. For example, what makes this technology unique includes:

– AI-powered autonomy. The drones can autonomously make decisions and adapt in real-time without needing human intervention, allowing for more efficient missions even in complex, dynamic environments. Quantum Systems says this was achieved by training the AI using deep reinforcement learning methods in a highly-specialised simulation environment. This allows the AI to refine its tactics through “continuous self-optimisation”, meaning it can make more efficient and precise decisions in tactical operations.

– Advanced sensor fusion. By integrating cutting-edge sensors and AI-based fusion algorithms, the swarms can gather and process a wide range of data (e.g. visual, infrared) in real-time, improving situational awareness and accuracy in missions.

– Collaborative behaviour. The swarms operate with a high level of coordination, thereby allowing the drones to perform complex tasks such as surveillance, search and rescue, or reconnaissance as a unified system, even when communication is limited or disrupted. As Quantum Systems says: “For the first time, a specially developed mission-AI controls and coordinates the UAS systems to ensure reliable mission execution even in scenarios with radio interference or a complete failure of individual drones”. 

– Scalability and flexibility. The system is designed to be scalable, enabling both large and small-scale drone swarms, and can be customised for diverse civilian and military applications, from disaster response to tactical operations.

What Happened In The Tests? 

In the successful tests, Quantum Systems says the Vector and Scorpion UAS from Quantum Systems and two other multi-purpose drones from Airbus were deployed in swarm flight and the reconnaissance data from all the drones was “merged in real time to form a joint situation picture and integrated into the Airbus ‘Fortion Joint C2’ battle management system”.  

Also, Quantum Systems has reported how the Vector drones demonstrated their ability to autonomously perform missions such as joint reconnaissance and target acquisition under GPS-denied conditions (GNSS denied), such as those found in Ukraine, thereby highlighting the ability of AI to increase the resilience of UAS to interference and ensure autonomous operation even under difficult conditions.

How Important Are Drones In Modern Warfare? 

Drones have become increasingly critical in modern warfare, particularly in conflicts like the ongoing war in Ukraine. In this war, both sides have used drones extensively for reconnaissance, surveillance, and strikes. For example, Ukraine deploys around 10,000 drones per month and relies heavily on smaller, commercial drones, such as DJI models, costing as little as $1,000 each. These drones are often repurposed with explosives for precision attacks.

The impact of drones in Ukraine is immense. They allow for real-time battlefield intelligence, enable faster response times, and reduce the cost of air operations. Also, the Ukrainian government has ramped up domestic drone production, with over 80 drone manufacturers now contributing to the war effort. As a result, drones have been a game-changer, levelling the playing field against larger forces, and are likely to dominate future conflicts.

What Next? 

Quantum Systems says the knowledge gained from its (KITU2) drone swarm study will help future developments to evaluate how learned behaviours from simulations can be integrated into real UAS systems, and the extent to which AI-controlled behaviours are superior to traditional manual control approaches. It also says, “The research results from the KITU2 study are intended to support the development of autonomous systems for major Bundeswehr projects such as the Main Ground Combat System (MGCS) and the Future Combat Air System (FCAS)”. 

As Sven Kruck, CRO and Managing Director, Quantum Systems says: “We are not just interested in expanding the technological capabilities of our drones. We want to give customers and users a real advantage in real-life scenarios. Ultimately, it’s about protecting soldiers and increasing safety. In the future, there will be no way around software-based and AI-supported systems for drone technology.” 

What Does This Mean For Your Business? 

The successful demonstration of Quantum Systems’ AI-powered drone swarm technology could be a significant milestone in modern military operations, because it shows how advancements in AI look like revolutionising the way drones are deployed in both military and civilian settings. With AI enabling autonomous decision-making and self-optimisation in real-time, this technology offers a more efficient and adaptive approach to complex missions, even in the most challenging conditions and could revolutionise modern warfare. Swarms of AI drones, learning and acting together, could have implications for future human deployment on the ground, perhaps taking us one small step further towards the idea of drone wars.

The potential applications of this technology also, thankfully, extend beyond warfare, offering promise in areas like disaster response, surveillance, and search and rescue. By demonstrating successful swarm coordination in highly dynamic and interference-prone environments, Quantum Systems has shown how AI can enhance the resilience, flexibility, and scalability of drone operations, setting the stage for future developments.

For other drone manufacturers and AI businesses, these advancements signal a growing demand for AI integration across drone systems. Companies that wish to stay competitive may now need to focus more on developing more intelligent and autonomous drones capable of performing complex tasks with minimal human input. The success of this technology opens up opportunities for collaboration between AI developers and industries such as agriculture, logistics, and defence, where advanced drone capabilities can be applied. This could also put pressure on AI firms to innovate further, particularly in the areas of machine learning algorithms, sensor fusion, and autonomous coordination, which will be increasingly critical as the industry moves towards smarter, more capable drone solutions.

As drones continue to play a pivotal role in modern conflicts, such as the ongoing war in Ukraine, and beyond into sectors like disaster management, the importance of AI-driven advancements can’t be overstated. Quantum Systems’ focus on integrating AI-learned behaviours into real-world systems, and its potential application to larger military projects like the Future Combat Air System (FCAS), highlights the transformative role AI looks likely to play in shaping the future of both military and civilian drone technology. This breakthrough reflects the broader trend towards AI-driven systems, and as these technologies evolve, they are poised to reshape industries far beyond the battlefield, offering new ways to manage national security and civilian crises.

New research compiled from AI-powered supply chain risk platform Infyos has revealed that 75 per cent of the lithium-ion battery supply chain may be linked to severe human rights abuses.

Human Rights Abuses – Forced (and Child) Labour 

Infyos’s analysis, which drew on government datasets, NGO reports, news articles, social media, and proprietary data, has revealed widespread human rights abuses in resource-rich countries where raw materials such as lithium and cobalt are mined and refined for lithium-ion batteries. These abuses, particularly involving forced and child labour, were found to be most prevalent in the early stages of the supply chain, notably during the extraction and processing of these critical materials.

Where? 

According to the analysis, much of this abuse appears to be concentrated in regions like Xinjiang, China, and countries with fragile governance, such as the Democratic Republic of Congo. In Xinjiang, allegations of forced labour are particularly severe, with accusations that many companies operating in the region are complicit. For example, it’s been suggested that companies that mine and refine lithium and cobalt in these regions may be involved in labour abuses, including instances where children as young as five are engaged in dangerous mining activities.

Link To The Battery Industry 

The demand for lithium-ion batteries has surged in recent years primarily due to the increased production of electric vehicles (EVs), the growth of renewable energy storage systems, and the expansion of portable electronic devices like smartphones and laptops. Governments and industries pushing for decarbonisation and net-zero emissions targets have also further driven this demand.

The battery industry’s connection to the alleged human rights abuses highlighted by Infyos stems from manufacturers sourcing components or materials from potentially unethical companies within their supply chain. These unethical practices are further obscured by complex business relationships, such as joint ventures or equity investments, where shifting ownership structures make it difficult to uncover the true extent of the exploitation.

As highlighted by Sarah Montgomery, CEO & Co-Founder, Infyos: “The relative opaqueness of battery supply chains and the complexity of supply chain legal requirements means current approaches like ESG audits are out of date and don’t comply with new regulations”. Sarah Montgomery added: “Most battery manufacturers and their customers, including automotive companies and grid-scale battery energy storage developers, still don’t have complete supply chain oversight.” 

So Many Suppliers 

One of the challenges that electric vehicle and battery manufacturers may have in identifying their supply chain risks is that they often have very complex supply chains, perhaps as many as 10,000 suppliers across their network, from mines to chemical refineries and automotive manufacturers. Human rights abuses upstream, e.g. at the raw materials stage (as identified by Infosys) may therefore be difficult to spot.

Not Just Infyos 

Infyos isn’t alone in suggesting human rights abuses in the lithium-ion battery supply chain. For example:

– Back in 2016, Amnesty International exposed child labour and hazardous working conditions in cobalt mining in the DRC, showing that some of the world’s largest electronics and automotive companies have not adequately addressed these risks.

– In 2023, the Business & Human Rights Resource Centre reported human rights violations and environmental damage related to lithium and cobalt mining in China, South America, and the DRC, with forced (and child) labour commonly involved.

– Also in 2023, Radio Free Asia reported uncovering human rights abuses and ecological damage in nickel mining in Indonesia and the Philippines, which provide critical materials for lithium-ion batteries, impacting local communities’ health and livelihoods.

Scrutiny 

However, the global battery supply chain is now under increasing scrutiny, particularly from regulators in Europe and the US. This is primarily due to growing concerns about human rights abuses such as forced (and child) labour in countries like the Democratic Republic of Congo (DRC) and China’s Xinjiang region. Legislation such as the EU Battery Regulation and the US Uyghur Forced Labour Prevention Act (UFLPA) are pushing companies to improve supply chain transparency and accountability. Non-compliance with these laws can result in products being blocked from key markets and heavy penalties, which could damage the reputation of the battery industry and slow down the energy transition. For example, companies are now at risk of losing investor confidence and facing financial penalties if they fail to manage these risks, with many already struggling to meet these stringent regulatory requirements.

What Can Be Done? 

To tackle these challenges, companies must adopt proactive measures to ensure ethical sourcing throughout their supply chains. This could include enhanced due diligence, where firms closely monitor their suppliers and implement robust Environmental, Social, and Governance (ESG) policies. Collaborating with independent auditors, utilising AI-based supply chain risk management tools like those provided by Infyos, and fostering stronger partnerships with suppliers may also be essential strategies. Also, companies must comply with emerging regulations, such as the battery passport system in the EU, which mandates rigorous supply chain traceability by 2027. By doing so, firms can not only avoid penalties but also align with investor expectations and contribute to a more sustainable future.

What Does This Mean For Your Business? 

With alternative battery types still some way off, as the demand for lithium-ion batteries continues to grow, so too does the urgency to address the human rights abuses linked to their supply chains. The findings from Infyos, alongside investigations by organisations like Amnesty International and the Business & Human Rights Resource Centre, serve as shocking reminders of the ethical complexities and the suffering behind these critical technologies. The global shift towards electric vehicles and renewable energy solutions must not be built on exploitation.

However, regulatory pressure is mounting, and companies that fail to ensure transparency and ethical sourcing will face significant reputational and financial risks. The path forward therefore does appear to be clear. By embracing stringent due diligence practices, enhancing supply chain visibility through AI-powered tools, and adhering to emerging regulations like the EU Battery Regulation, the industry can foster a more responsible and sustainable future. That said, in the real world, many companies may be deterred by the high costs of implementing such measures, especially in complex global supply chains. The vastness and opacity of these networks, coupled with competitive pressures to keep costs low, may make ethical sourcing less of a priority. Also, inconsistent enforcement of regulations and varying levels of consumer concern about supply chain ethics could further reduce the incentive for businesses to fully embrace transparency and accountability that’s needed.

Ultimately, the energy transition depends not only on technological innovation but also on a commitment to human rights and ethical practices. For the battery industry to truly support a greener future, it must first ensure that its foundations are just and free from exploitation.

Following its recent release, Apple’s latest macOS update, dubbed Sequoia (macOS 15), has disrupted the functionality of several widely used cybersecurity tools, including those from CrowdStrike, SentinelOne, and Microsoft.

Users and developers have voiced frustrations on social media and in Mac-focused forums about issues leaving many security applications non-operational. Reports highlight problems with tools like CrowdStrike, SentinelOne, Microsoft Defender, and ESET, alongside browser issues, particularly with Firefox, where the OS firewall sometimes blocks web access. The root cause is unclear, but the disruptions are creating significant challenges for both end-users and enterprise security teams.

The key issue seems to stem from changes Apple has made in Sequoia’s network stack, which cybersecurity firms say is interfering with their products. For example, CrowdStrike delayed support for Sequoia, citing complications in adapting their software. Similarly, SentinelOne warned users not to upgrade without ensuring proper support, while Microsoft’s Defender and ESET have faced similar difficulties. Despite SentinelOne and ESET eventually providing compatibility, significant disruption remains across the community.

Apple has yet to comment, leaving security firms and users to manage the situation. CrowdStrike is awaiting a Sequoia update to offer full support, while temporary workarounds are being shared to address issues like firewall settings and basic web browsing. For now, it may be advisable to delay upgrading to macOS Sequoia until Apple or security providers release compatible updates.

Google has announced that users can now securely sync passkeys across all devices, not just Android, making sign-ins faster and more secure.

Passkeys use biometrics, such as your fingerprint, face, or screen lock, to sign in to apps and websites, thereby making it easier and more secure than using traditional passwords. Whereas previously, only Android devices could save passkeys, requiring a QR code scan for use elsewhere, Google’s latest update means users can now save passkeys to Google Password Manager on desktops (Windows, macOS, and Linux), with ChromeOS in Beta.

Once a passkey is saved, it syncs automatically across devices. Google has also introduced a Google Password Manager PIN for extra security, ensuring passkeys are encrypted and inaccessible to others, even Google.

To start using use passkeys on a new device, users will need their Google Password Manager PIN or Android screen lock. Users can set up a six-digit recovery PIN by default or select “PIN options” to create a longer alpha-numeric PIN.

Google says that with passkeys available for sites like Google, Amazon, and PayPal, and with Google Password Manager built into Chrome and Android, users can start benefitting from this secure, more convenient change without the need for extra apps.

Engineers at the Massachusetts Institute of Technology (MIT) are developing a new kind of reconfigurable masonry made from 3D-printed, recycled glass.

Fits Together Like LEGO 

MIT says the new multilayered glass bricks, each in the shape of a figure-of-eight, are designed to interlock, much like LEGO bricks.

3D Printed 

One of the big advantages of the new glass bricks is that they are made using a custom 3D glass printing technology (provided by MIT spinoff Evenline). The inspiration for using glass and the brick’s shape came partly from when 2 of the engineers, Kaitlyn Becker, and Michael Stern, were still undergraduates and learned the art and science of blowing glass in MIT’s Glass Lab.

It was this experience that led Stern to design a 3D printer capable of printing molten recycled glass.

Tested 

Becker and Stern collaborated to test whether 3D-printed glass could function as structural masonry units comparable to traditional bricks. Using the latest version of Evenline’s Glass 3D Printer (G3DP3), which melts recycled glass bottles into a printable form, they produced prototype bricks from soda-lime glass. The figure-eight design bricks featured two round pegs, similar to LEGO studs, allowing them to interlock and form larger structures. A removable material between bricks prevented scratches, enabling easy dismantling and recycling.

Strong 

The MIT team tested the glass bricks’ strength using an industrial hydraulic press and found that the strongest bricks could withstand pressures similar to concrete blocks. These bricks were primarily made of printed glass, with a separately manufactured interlocking feature, suggesting that most of the brick can be printed from glass, while the interlocking part can be made from various materials.

The Advantages 

The many advantages of the 3D-printed glass brick system include:

– Sustainability. The bricks are made from recycled glass, supporting circular construction by reusing materials and reducing the need for new manufacturing, which lowers the construction industry’s embodied carbon.

– Reusability. The bricks can be disassembled and reassembled multiple times for different structures, extending their lifespan across generations of buildings.

– Recyclability. Glass is highly recyclable. For example, the glass used by MIT’s 3D printer comes primarily from recycled glass bottles in the first place which are crushed, melted in a furnace, and then transformed into a molten, printable material used in the 3D glass printer. Also, once the glass bricks have been made, they can be remelted again and reshaped without contamination, allowing bricks to be recycled into new forms. As Kaitlyn Becker, assistant professor of mechanical engineering at MIT says: “We’re taking glass and turning it into masonry that, at the end of a structure’s life, can be disassembled and reassembled into a new structure, or can be stuck back into the printer and turned into a completely different shape. All this builds into our idea of a sustainable, circular building material.” Becker also highlights how, “As long as it’s not contaminated, you can recycle glass almost infinitely”.

– Strength. As highlighted in MIT’s mechanical tests, the glass bricks can withstand pressures similar to concrete, making them viable for structural use.

– Their interlocking design. Like LEGO, the bricks feature interlocking pegs, enabling easy assembly, and creating strong, self-supporting structures.

– Scratch and crack prevention. A removable material between bricks prevents damage during assembly and dismantling.

– Adaptability. The figure-of-eight design allows for curved wall constructions and offers flexibility in design. This allows for more creative and varied structural forms, making it possible to create aesthetically unique and functional buildings that traditional brick designs may not easily support.

– The potential for scalability. The system can be scaled up to create larger structures, with potential for various configurations and reconfigurations. As Stern says: “We have more understanding of what the material’s limits are, and how to scale,” and that “We’re thinking of stepping stones to buildings, and want to start with something like a pavilion – a temporary structure that humans can interact with, and that you could then reconfigure into a second design.” 

– The environmental benefit of minimising the manufacturing of new materials and reducing the construction industry’s “embodied carbon”, i.e. the greenhouse gas emissions associated with every process throughout a building’s construction, from manufacturing to demolition.

Drawbacks? 

Although the system is still at the development stage and the engineers have been keen to highlight the advantages of the system, it is possible to think of some of the more obvious potential disadvantages, such as:

– Producing glass bricks using 3D printing requires specialised equipment and processes, which might be more expensive and complex than traditional brickmaking.

– Glass typically has poor insulating properties, so structures made from glass bricks may not retain heat as effectively as those built with traditional materials.

– The need for a separate interlocking feature made from a different material could complicate the production and assembly process, reducing the system’s simplicity and uniformity.

– Widespread use of glass bricks might face resistance due to unfamiliarity or scepticism about their long-term durability and safety in construction. Also, the unusual shape and the fact that it’s a new material may require training, e.g. for builders.

Glass Already Being Used To Make Bricks 

Although the 3D printer idea for full glass bricks is new, it’s worth noting here that recycled glass is already being experimented with in similar ways for use in construction projects. For example, researchers at Nanyang Technological University (NTU) in Singapore have developed a concrete mix using recycled glass as a substitute for sand, which is increasingly scarce due to overuse. This glass-based concrete has been successfully used in 3D printing to create a 40 cm-tall concrete bench, demonstrating its viability for load-bearing construction applications.

What Does This Mean For Your Organisation? 

The development of 3D-printed glass bricks at MIT presents a promising and bold vision for sustainable construction, combining innovation in design and environmental responsibility. By reimagining glass as a structural material and leveraging 3D printing technology, these interlocking bricks could offer a versatile solution that embraces circular construction principles. As the building industry seeks to reduce its environmental impact, these bricks present a potential alternative by utilising recycled glass, minimising waste, and allowing structures to be easily reconfigured and recycled at the end of their lifespan.

While challenges remain, such as higher production costs and concerns about insulation and durability, the adaptability and recyclability of the glass bricks highlight their potential.  As with any new material and system though, acceptance and implementation are likely to take time, something that we’re running out of when it comes to decarbonising industries.

However, looking on the bright side, the demonstrated strength of the bricks, combined with their aesthetic and sustainable benefits, points towards a future where glass could play a significant role in eco-friendly construction. The success of this system could even pave the way for further exploration of recycled materials in 3D printing, and with continued innovation, it’s possible to see how these glass bricks and/or concrete using crushed up glass instead of sand, could become a cornerstone in the move towards more sustainable building practices.

This video tutorial explains in depth how to use ChatGPT to optimise your LinkedIn profile.

[Note – To Watch This Video without glitches/interruptions, It’s best to download it first]

The Magnifier tool allows you to zoom in on any part of your screen, which is especially useful when working with small text or detailed images during presentations or document reviews. Here’s how it works:

Enable Magnifier

– Press Win + Plus (+) to activate the Magnifier.

Adjust Zoom Levels

– Use Win + Plus (+) to zoom in and Win + Minus (-) to zoom out.

Exit Magnifier

– Press Win + Esc to turn off Magnifier.

ChatGPT’s maker, OpenAI, has announced the introduction of its new OpenAI o1 large language model that can use “complex reasoning” to fact-check its own answers before giving them.

ChatGPT Plus and Teams Users Can Try It Now 

The new o1 model is already available to ChatGPT Plus or Team users, and in OpenAI’s API. OpenAI o1 is a series of AI models, currently comprising of a ‘Preview’ version, which Open says uses “advanced reasoning”, or the o1-mini (lighter version), which OpenAI says is “Faster at reasoning” (than its other models), and is particularly good for coding tasks.

What’s So Different About o1? 

What sets OpenAI o1 apart from other models is its enhanced reasoning capabilities, designed to tackle complex, multi-step problems with a thoughtful, more holistic approach. Unlike previous models like GPT-4 (which focus on speed), o1 takes more time to “think through” problems, improving its performance on tasks requiring deeper analysis, such as advanced coding, mathematical reasoning, and document comparison. OpenAI says this is because o1 has been “trained with reinforcement learning to perform complex reasoning. o1 thinks before it answers – it can produce a long internal chain of thought before responding to the user”. 

OpenAI says it’s the reinforcement learning (train-time compute) and the extra time ‘thinking’ (test-time compute) that significantly reduces hallucinations. In other words, it is sacrificing speed in favour of accuracy and depth , enabling o1 to excel at complex problem-solving. This should make o1 ideal for use cases where precision is more critical than quick responses. ChatGPT users may argue, however, that precision, i.e. real (not ‘made-up’ answers) has always been completely necessary.

Chain-of-Thought Approach 

OpenAI says the fact that o1 uses a chain-of-thought when attempting to solve a problem, thinking for a long time before responding to a difficult question as humans do, is a large part of the secret of its apparent success. The fact that o1 breaks down “tricky steps into simpler ones” and “learns to try a different approach when the current one isn’t working” is credited with being the process that “dramatically improves the model’s ability to reason”. 

Just How Good Is It? 

OpenAI says that to highlight the reasoning improvement over GPT-4o, it tested the o1 models on a diverse set of human exams and ML benchmarks. For example, to test o1 on chemistry, physics, and biology, OpenAI used the GPQA diamond, a difficult intelligence benchmark in those subjects, recruited experts with PhDs to answer GPQA-diamond questions, and compared o1’s answers with theirs. For mathematics, OpenAI evaluated o1’s performance on AIME, an exam designed to challenge the brightest high school math students in America. Also, in coding, OpenAI simulated competitive programming contests hosted by Codeforces to demonstrate o1’s coding skill.

The results are reported to show that o1 “ranks in the 89th percentile on competitive programming questions (Codeforces), places among the top 500 students in the US in a qualifier for the USA Math Olympiad (AIME) and exceeds human PhD-level accuracy on a benchmark of physics, biology, and chemistry problems (GPQA)”.   

In short, OpenAI says: “In many reasoning-heavy benchmarks, o1 rivals the performance of human experts”. 

Not Ideal For All Use Cases 

The new o1 was also evaluated in terms of human preference – i.e. what people found it was best for (compared to OpenAI’s other models). The o1-Preview was preferred over GPT-4o in reasoning-heavy tasks like data analysis, coding, and maths, due to its advanced problem-solving capabilities. However, it was less favoured for certain natural language tasks, indicating that while it excels in technical reasoning, it may not be ideal for all types of use cases.

Seems Safer 

OpenAI says the ‘chain-of-thought’ reasoning (outlined earlier) in o1 Preview helps improve safety by integrating human values and safety rules into its decision-making process (the model has been taught OpenAi’s safety rules and how to reason about them in context), thereby making the model more robust and effective in refusing unsafe requests. The chain-of-thought approach is also beneficial because it enables users to see the model’s reasoning process – users can observe the model’s thinking in a legible way, and it ensures better handling of unexpected situations, especially in sensitive tasks. For users of o1, this could mean increased reliability and trustworthiness, especially in environments where safety and ethical concerns are critical.

Illustration 

The importance of safety in AI models was recently illustrated by an artist and hacker (working under the name ‘Amadon’), who reported how he was able to fool ChatGPT into ignoring its own guidelines and ethical responsibilities to provide him with instructions for making powerful explosives! Amadon reportedly described his process as a “social engineering hack to completely break all the guardrails around ChatGPT’s output.” 

In an operation known as a “jailbreak” (i.e., tricking a chatbot into operating outside of its preprogrammed restrictions), Amadon reportedly told ChatGPT to give him the bomb-making instructions by telling the bot to “play a game”. He then followed this up with more, related prompts with the intention of creating a fantasy world where the real rules and guidelines of the chatbot would no longer apply.

This is worrying because it demonstrates how even advanced AI systems are vulnerable to being manipulated to perform potentially dangerous tasks. This could mean that individuals with malicious intent could exploit such vulnerabilities, compromising public safety and undermining trust in AI’s ethical boundaries. Let’s hope o1 can use its chain-of-thought approach to take the time to realise it’s being fooled and deliver a well-thought-out ‘no’ to anyone who tries to jailbreak it.

Other Disadvantages 

Other apparent disadvantages of o1 (from what can be seen so far) include:

– Its ‘basic’ functionality, i.e. it currently lacks key features such as web browsing and file analysis, and its image analysing capabilities are temporarily pending further testing.

– Users are restricted by weekly message limits. For example, o1-Preview allows 30 messages and o1-mini is capped at 50 messages.

– It’s relatively expensive. o1-Preview is priced at $15 per million input tokens and $60 per million output tokens, meaning it’s significantly more expensive than GPT-4o.

Competitor – Google 

OpenAI isn’t the only company developing reasoning methods in its models. Google DeepMind’s AlphaProof and AlphaGeometry 2, for example, have shown remarkable progress in mathematical reasoning. These models were trained using formal languages to solve high-level maths problems, as seen in their performance at the 2024 International Mathematical Olympiad (IMO). AlphaProof uses reinforcement learning to verify mathematical proofs, enabling it to tackle increasingly complex problems. This emphasis on formalised reasoning sets it apart from OpenAI’s more general-purpose approach.

What Does This Mean For Your Business? 

The introduction of OpenAI’s o1 model could have significant implications for businesses looking to adopt generative AI. For businesses that need accuracy and reliability, particularly in fields requiring complex reasoning such as data analysis, coding, or scientific problem-solving, o1 appears to offer a solution that could dramatically enhance productivity. The model’s chain-of-thought reasoning makes it more capable of reducing errors and providing accurate outputs, making it ideal for industries where precision is essential.

For OpenAI, the launch of o1 helps it differentiate itself from competitors, such as Google DeepMind, by focusing on general-purpose reasoning and problem-solving rather than highly specialised tasks (although o1-mini is supposed to be particularly good at coding). However, the slower inference time and higher costs may deter businesses seeking faster, more cost-efficient solutions for simpler tasks. This could leave room for competitors to attract users who require speed and versatility rather than deep analytical capabilities.

For business users, o1 does appear to present an opportunity to integrate a more reliable and safe AI system, especially important for industries dealing with sensitive data and complex decision-making. Yet, its higher price and current lack of key functionalities like web browsing or file analysis mean that businesses must carefully evaluate if o1 aligns with their specific needs. Trust and efficiency are crucial for businesses adopting AI, and while o1 excels in reasoning-heavy applications, organisations will need to balance its current strengths against these limitations when considering whether or when to implement it.