UK banking customers who have faced repeated IT failures are set to receive millions in compensation, following a damning Treasury Committee investigation into the scale and impact of these outages.

Nine Banks

The inquiry revealed that in the past two years alone, nine major banks and building societies have suffered more than a month’s worth of system failures, leaving millions unable to access their money when they needed it most.

How Big Is The Problem?

The inquiry revealed that between January 2023 and February 2025, major high street banks (including Barclays, HSBC, Lloyds, Nationwide, Santander, and NatWest) collectively experienced at least 803 hours of IT outages. That’s more than 33 days of service disruptions, affecting customers’ ability to make payments, transfer funds, and in some cases, even access their own accounts.

To make matters worse, some of the most disruptive outages occurred on key dates, including payday, adding to the distress caused. For many, these incidents resulted in late bill payments, missed wages, and even the inability to complete major transactions – a situation that has been described as ‘deeply unsettling’ by campaigners.

Many Living Pay Cheque to Pay Cheque

Commenting on the findings of the report, Dame Meg Hillier, Chair of the Treasury Committee, did not hold back in her criticism of the situation, stating: “For families and individuals living pay cheque to pay cheque, losing access to banking services on payday can be a terrifying experience. The fact there has been enough outages to fill a whole month within the last two years shows customers’ frustrations are completely valid.”

Which Banks Are Paying Out and How Much?

As scrutiny on the sector intensifies, it seems that some banks have now set aside millions to compensate affected customers. Here’s what each bank is paying and why:

– Barclays… The worst-hit bank in the report, Barclays is expected to pay between £5 million and £7.5 million for inconvenience and distress caused by various outages, with the total amounting to £12.5 million over the past two years.

– Bank of Ireland… The second-largest compensation payout, with £350,000 being distributed to impacted customers.

– NatWest… Has recorded 13 major incidents and will be compensating customers £348,000.

– HSBC… With 32 separate IT failures, HSBC has set aside £232,697 in compensation payments.

– Lloyds… Customers will receive £160,000 following 12 incidents of service disruption.

– Nationwide… The building society has paid out £77,452 due to system failures.

– Santander… Despite 24 incidents, its compensation stands at just £17,000.

– AIB… Paying out a nominal £590 in compensation.

The scale of Barclays’ payout alone shows the gravity of the problem, particularly given that its most recent failure in January left 56 per cent of online payments failing on payday, with some customers unable to complete house moves and others left stranded without funds.

What It Could Mean for Customers

For banking customers, these payouts are likely to come as both a relief and a frustration. For example, while the compensation acknowledges the distress caused, it’s likely to do little to restore confidence in the reliability of banking services. Many customers have already expressed concerns that IT failures are becoming more frequent, rather than less, despite advances in technology.

Although the payouts may be welcomed, customers will still be aware that unless something is done about the failing IT systems from underlying legacy banking infrastructure that keeps crashing, they’re still at risk of it happening again.

Unclear How to Claim Compensation

While Barclays and others have vowed that no customer will be left out of pocket, the process of claiming compensation still remains unclear for many. Some have already lodged formal complaints, while others may need to wait for banks to proactively contact them about payments.

The Wider Impact on the Banking Industry

This wave of IT failures, and the subsequent compensation payouts, has put the UK banking industry under renewed pressure to modernise its digital infrastructure. Thankfully, the Treasury Committee has made it clear that more needs to be done to reduce the frequency of these failures, and banks are now being urged to make urgent investments in:

– IT resilience – ensuring systems are robust enough to handle peak usage periods.

– Third-party oversight – many failures stem from external suppliers, raising questions about regulation.

– Customer communication – better transparency when outages occur and clearer processes for compensation.

The government has already hinted that further regulation may be on the horizon, with a Treasury spokesperson stating: “We are working with the financial authorities to regulate third-party suppliers, as well as considering whether the banks are doing all they can to provide the level of service customers expect.”

What Does This Mean for Your Business?

The banks now find themselves at a crossroads. While the compensation payments acknowledge the impact on customers, they do not solve the deeper issue of unreliable IT infrastructure. For customers, the payouts may soften the blow, but they will do little to restore long-term confidence unless meaningful action is taken to prevent future disruptions.

For businesses, the implications of these failures are always far-reaching. For example, many companies rely on seamless banking operations to pay staff, manage cash flow, and complete critical transactions. Therefore, when banking systems go down, the knock-on effects can be severe, potentially leading to delayed wages, operational disruptions, and financial uncertainty. Small businesses, in particular, can struggle to absorb these setbacks, making banking reliability an essential component of economic stability.

The industry’s response to this crisis will shape the future of banking in the UK. If banks commit to modernising their IT systems and prioritising customer service, they may yet rebuild trust. However, continued failures could prompt stricter regulatory intervention, higher penalties, and increased competition from digital-only banks that have so far proven more resilient. With technology at the heart of modern finance, institutions that fail to adapt may find themselves losing not just customers, but also their position in the market.

As pressure mounts, it seems that banking in the UK is at a kind of turning point, and the coming months will determine whether these institutions can step up to meet the expectations of the businesses and individuals who depend on them every day.

In this Tech Insight, we look at the latest AI developments, including AI agents creating their own language, Opera’s new AI-powered browsing assistant, and a growing debate over the risks of an international race to artificial general intelligence.

AI Agents Are Speaking Their Own Language

A new experiment called ‘GibberLink’ has just demonstrated an intriguing concept, i.e. AI voice agents that can recognise when they are speaking to another AI and then switch to a more efficient communication method (which is incomprehensible to humans). Developed at a hackathon in London by Meta engineers Boris Starkov and Anton Pidkuiko, GibberLink replaces human-like speech with GGWave, a sound-based protocol that allows AI systems to exchange information faster and with less computing power.

To human ears, the communication sounds like a series of beeps and boops, reminiscent of dial-up internet modems from the 1980s. While it may seem like a niche experiment, the technology could actually have real-world applications. For example, as companies increasingly deploy AI-powered customer service agents, there may be a need for them to communicate directly. By using this machine-native ‘language,’ AI agents could cut computing costs significantly, making AI-driven voice interactions cheaper and more efficient.

The apparent viral success of GibberLink has sparked both fascination and concern. For example, some fear that AI developing its own incomprehensible communication methods could reduce transparency and accountability. However, Starkov and Pidkuiko insist their project is simply an experiment, and they have open-sourced the code rather than commercialising it (at least for now).

Opera’s AI Assistant Could Change How We Browse

It seems that Opera, the long-standing web browser, is aiming to redefine internet browsing with its AI-powered feature called ‘Browser Operator’. Unlike traditional AI assistants that simply summarise search results, Browser Operator actively completes tasks for users. For example, if you need to book a flight, it’s just a case of providing a few details, and the AI will search, compare prices, and add the best option for you.

What sets Opera’s AI apart from competitors like OpenAI’s Operator and Anthropic’s Claude is that it operates locally within the browser. This ensures not only faster performance but also better privacy, as user data never leaves the device. Unlike cloud-based AI tools, which require remote servers to function, Browser Operator processes requests directly on the user’s machine.

The potential benefits are clear, i.e. more efficient browsing, time saved on repetitive online tasks, and a more seamless digital experience. However, as with any AI-driven automation, questions remain about how much control users will ultimately have over decisions made on their behalf. Opera has included safeguards, allowing users to pause or cancel tasks at any time, ensuring that humans remain in the loop.

Eric Schmidt Warns Against an AI Arms Race

Amidst the rapid AI advancements, former Google CEO Eric Schmidt and two co-authors, Alexandr Wang and Dan Hendrycks, have just published a policy paper warning against a US-led push for artificial general intelligence (AGI). It seems they’re concerned that an aggressive AI race could provoke international retaliation, particularly from China, and potentially destabilise global relations.

Schmidt and his colleagues argue that the pursuit of AGI (a hypothetical AI system with intelligence surpassing human capabilities) should not be treated like the Manhattan Project, the US government’s programme to develop nuclear weapons in the 1940s. Instead, they propose a more cautious strategy, advocating for ‘Mutual Assured AI Malfunction’ (MAIM), which suggests that governments should focus on defensive measures to deter AI-driven threats rather than escalating an AI arms race.

The paper challenges the notion that the US must ‘win’ the AGI race, arguing that such thinking could lead to dangerous consequences, including pre-emptive cyberattacks from adversaries. Instead, they suggest limiting access to powerful AI chips, strengthening cybersecurity, and ensuring AI remains under human control.

What Does This Mean for Your Business?

These developments show just how rapidly AI is reshaping industries, and why businesses need to stay ahead of the curve. The rise of AI agents like GibberLink suggests a future where automated systems could interact more efficiently without human intervention. While this might reduce costs, businesses relying on AI-driven communication must consider transparency and ethical oversight to maintain customer trust and regulatory compliance.

Meanwhile, Opera’s Browser Operator signals a shift in how AI can automate everyday digital tasks. For businesses, this could mean exploring ways to integrate AI-powered automation into their workflows to improve efficiency. Whether it’s customer service, e-commerce, or operations, AI-driven task completion could actually free up human employees for more strategic work. However, as with any AI system, companies will need to carefully manage data privacy concerns and ensure these tools remain user-controlled rather than fully autonomous.

Also, the debate over AGI development highlights broader implications for businesses investing in AI research. If governments take a more defensive stance, regulations around AI chips and open-source models could tighten, limiting access to cutting-edge AI innovations. For businesses in cybersecurity, cloud computing, and AI ethics, this could create new opportunities (but also new risks). Understanding the shifting regulatory landscape will be critical for companies looking to leverage AI without falling foul of future legal constraints.

The main message here is, therefore, that as AI continues to evolve, businesses that can embrace its efficiencies while maintaining ethical and regulatory oversight are likely to be best positioned for long-term success. Whether integrating AI assistants, automating customer interactions, or staying informed on global AI policy, companies that adapt strategically will stay competitive in an increasingly AI-driven world.

According to a report by The Information, OpenAI is gearing up to introduce a new wave of specialised ‘AI agents’, with some of its most advanced offerings set to cost as much as $20,000 per month.

New Revenue Stream For OpenAI

With AI development costs rising fast, OpenAI reportedly lost around $5 billion last year due to the huge expense of running and improving its services. It seems, therefore, that facing growing competition, OpenAI may now be looking for new revenue streams, such as the high-end AI agents idea.

What Are OpenAI’s AI Agents?

AI agents are essentially advanced, self-running AI systems that go beyond simple prompts. Unlike standard AI models that require users to request information, OpenAI’s planned AI agents will be more autonomous. For example, they won’t just answer questions but will proactively analyse data, generate insights, and even carry out complex workflows independently.

Tiers of AI Agents

Reports from The Information suggest that OpenAI may actually offer several tiers of AI agents, each designed for different professional applications. For example:

– High-Income Knowledge Worker Agent – This entry-level agent, priced at $2,000 per month, is intended to assist professionals in managing and interpreting large volumes of information.

– Software Developer Agent – At $10,000 per month, this agent will focus on automating coding tasks, debugging, and optimising software development processes.

– PhD-Level Research Agent – The most advanced and expensive option, priced at $20,000 per month, is designed for high-level research in academia, science, and other knowledge-intensive fields.

Why Is OpenAI Introducing These High-Cost AI Agents?

The pricing structure for these agents reflects both the increasing demand for advanced AI-driven automation and the significant costs associated with developing and running such systems. For example, OpenAI’s AI models require enormous computing power and continuous updates, making them expensive to maintain.

Also, as mentioned earlier, OpenAI is facing mounting pressure to become financially sustainable. Some commentators have suggested that despite its rapid growth, the company’s costs are continuing to outstrip revenue, thereby leading OpenAI to look for high-value enterprise customers who can afford premium AI solutions.

When Could These AI Agents Launch?

While OpenAI has not officially confirmed a launch date, The Information reports that the company is already in active discussions with potential customers. Some reports indicate that businesses may start accessing these agents within the next few months, with a full-scale rollout possibly occurring by the end of the year.

OpenAI has already secured some significant financial backing to support its AI agent plans. For example, The Information reported that SoftBank, a major investor in the AI firm, has committed to spending $3 billion on OpenAI’s agent products this year alone.

Who Will Use These AI Agents?

Given the hefty price tag, OpenAI’s AI agents are unlikely to be aimed at individual users or small businesses. Instead, they will most likely be adopted by large corporations, research institutions, and high-net-worth professionals who require AI-driven expertise in complex fields.

Examples of sectors that could benefit from these AI agents include:

Academic and Scientific Research – AI agents could assist in processing vast amounts of data, identifying patterns, and even generating new theories.

Finance and Investment – Hedge funds and financial analysts may use AI agents for risk assessment, forecasting, and automating trading strategies.

Technology and Software Development – AI-powered software engineering agents could help accelerate innovation and optimise complex coding tasks.

Legal and Corporate Advisory Services – Law firms and consultants could use AI agents for research, contract analysis, and regulatory compliance monitoring.

The Effect of OpenAI’s Agent Pricing Strategy

The effects of OpenAI’s introduction of these high-priced AI agents could include:

– AI as an Elite Tool – At $20,000 per month, access to the most advanced AI capabilities could be restricted to large organisations and well-funded institutions, potentially increasing disparities in AI adoption.

– New Business Models in AI – OpenAI’s move could set a precedent for other AI firms to introduce premium AI services, shifting AI from a general consumer tool to a specialised enterprise product.

– Pressure on Competitors – Companies like Google DeepMind and Anthropic may be forced to introduce competing AI products, potentially driving innovation but also raising concerns about escalating AI costs.

– Regulatory Considerations – As AI agents become more autonomous, regulators may need to step in to establish guidelines for their deployment and use.

OpenAI’s Financial Challenges and Strategic Partnerships

OpenAI’s decision to introduce premium AI agents comes amid efforts to strengthen its financial position. According to The Information’s report, OpenAI aims to raise up to $40 billion in a new funding round, which could push its valuation to as high as $300 billion! SoftBank is expected to play a key role in this funding round, with reports suggesting it could invest between $15 billion and $25 billion.

However, despite these efforts, OpenAI is still facing long-term financial risks. For example, analysts predict the company could lose up to $44 billion before achieving profitability, with its computing expenses potentially rising to $37.5 billion annually by 2029. These financial pressures may force OpenAI to further increase prices or seek new monetisation strategies in the future.

What Does This Mean For Your Business?

For UK businesses, what these tiers of high-price AI agents could essentially mean is the development of a growing divide between those able to leverage AI-driven efficiencies and those left behind due to cost barriers. High-end AI could, therefore, be moving firmly into the realm of enterprise solutions rather than everyday consumer use. For example, large corporations and institutions may see these tools as a way to cut costs and boost productivity, but smaller businesses could struggle to compete if access to advanced AI remains financially out of reach.

This could mean that the AI industry itself may also shift towards more premium, subscription-based models, where access to top-tier AI capabilities is increasingly restricted to those with the deepest pockets. Competitors such as Google DeepMind and Anthropic may be forced to adjust their pricing strategies or innovate further to stay competitive. Also, regulators may step-in to ensure AI developments remain accessible and ethically governed.

OpenAI’s decision to pursue high-end enterprise customers could also be seen as being part of a broader trend in AI commercialisation, where profitability and sustainability are now as much a priority as innovation. While AI remains a rapidly evolving field, there is increasing pressure on leading AI companies to justify their valuations and revenue models. With OpenAI reportedly seeking to raise billions in new investment while tackling significant losses, the success of its AI agents could help determine whether this is a long-term strategic shift or just a short-term response to financial pressures. The wider AI industry will, no doubt, be watching closely to see if this signals a sustainable future for advanced AI or simply a new phase in an already volatile market.

Non-profit web tech organisation Mozilla has rewritten Firefox’s Terms of Use after widespread criticism from users who feared the company was overreaching with its data policies.

Vague Wording?

It seems that the backlash stemmed from vague and broad wording in the updated terms, which many believed gave Mozilla rights over any data entered into the browser.

After days of mounting pressure, Mozilla admitted the wording had caused unnecessary concern and moved quickly to clarify its position. The rewritten terms now explicitly state that Mozilla does not claim ownership over user data and is only processing it to operate Firefox as intended.

How the Controversy Started

The trouble began when Mozilla introduced its updated Terms of Use and Privacy Notice for Firefox users in late February. The update was meant to improve transparency, but instead, it sparked alarm. The particular clause in the new terms that became the centre of the controversy was:

“When you upload or input information through Firefox, you hereby grant us a non-exclusive, royalty-free, worldwide licence to use that information to help you navigate, experience, and interact with online content as you indicate with your use of Firefox.”

Unfortunately for Mozilla, many users appeared to interpret this as Mozilla giving itself sweeping rights over everything they typed or uploaded through Firefox, thereby raising concerns about potential data sales to advertisers or AI companies.

To make matters worse, Mozilla had also removed a previous assurance from its Privacy FAQ that explicitly stated the company would never sell user data. This change further fuelled speculation that Mozilla was shifting its stance on privacy.

Mozilla’s Swift Rewrite

The backlash was immediate, with users taking to forums, social media, and tech news websites to voice their concerns. It seems that many long-time Firefox supporters, who valued Mozilla’s strong stance on privacy, may have felt betrayed.

Realising the scale of the problem, Mozilla responded swiftly. Ajit Varma, Mozilla’s Vice President of Product, acknowledged the confusion, stating: “Our intent was just to be as clear as possible about how we make Firefox work, but in doing so we also created some confusion and concern.”

To address these concerns, Mozilla rewrote the contentious section of the Terms of Use. The revised wording now states:

“You give Mozilla the rights necessary to operate Firefox. This includes processing your data as we describe in the Firefox Privacy Notice. It also includes a non-exclusive, royalty-free, worldwide licence for the purpose of doing as you request with the content you input in Firefox. This does not give Mozilla any ownership in that content.”

In addition to clarifying its stance on data ownership, Mozilla also removed a reference to its Acceptable Use Policy, which had caused further confusion.

Why Did Mozilla Change the Terms in the First Place?

The controversy wasn’t caused by a sudden change in how Mozilla handles data, but rather by the legal complexities surrounding privacy laws.

Mozilla explained that it had stepped away from making blanket claims like “We never sell your data” because privacy laws in different jurisdictions define “selling” in different ways. For example, the California Consumer Privacy Act (CCPA) considers “selling” to include actions like “disclosing, making available, or transferring data,” even if no money changes hands.

Given these evolving legal definitions, Mozilla had adjusted its wording to avoid potential legal risks. However, in doing so, it unintentionally created alarm among users who feared a change in its long-standing commitment to privacy.

Does Firefox Actually Share User Data?

Despite the backlash, Mozilla insists that its approach to data privacy has not changed. The company has stated that yes, Firefox does collect and share some data, but it is done in a way that protects user privacy.

Mozilla says it doesn’t track individual browsing activity or sell personally identifiable information. Instead, it collects aggregated and anonymised data to support:

– Optional ads on the New Tab page

– Sponsored suggestions in the search bar

– Performance tracking to improve Firefox’s functionality.

Mozilla stresses that users still have full control over their data-sharing settings, which can be customised in Firefox’s privacy controls.

What Does This Mean For Your Business?

While Mozilla acted quickly to correct the situation, this whole episode appears to have damaged trust among some users. The controversy highlights how even a company with a strong privacy reputation can stumble when communication isn’t perhaps as clear as it could be.

For privacy-conscious users, such as businesses, the backlash should serve as a reminder to remain vigilant about changes to terms and privacy policies, even from companies that market themselves as pro-privacy.

For Mozilla, this incident underlines the need for better communication when handling sensitive topics like user data. It also shows how legal wording, even if well-intended, can backfire if it leaves room for misinterpretation.

With Firefox struggling to compete against dominant browsers like Google Chrome, maintaining trust is crucial for Mozilla’s future. The company may have resolved this issue for now, but it will need to tread carefully to avoid similar missteps in the future.

Brother (once considered one of the more consumer-friendly printer manufacturers) is now under fire over claims that its latest firmware updates deliberately degrade print quality when third-party toner cartridges are used.

Firmware Updates Allegedly Restrict Third-Party Ink

Right-to-repair activist and electronics technician Louis Rossmann has accused Brother of deploying firmware updates that hinder the use of non-OEM (Original Equipment Manufacturer) consumables. According to Rossmann, users have reported that certain printer functions, such as automatic colour registration, stop working after an update when a third-party cartridge is detected.

Several users have taken to forums and social media to confirm similar issues, with reports stating that previously functional non-OEM toner cartridges suddenly started producing lower-quality prints or experiencing feature restrictions. Furthermore, rolling back to an earlier firmware version is proving difficult, as Brother has removed older updates from its servers.

Brother Denies Any Wrongdoing

Brother has responded to the allegations, denying any deliberate attempt to block third-party consumables. The company issued a statement saying:

“We are aware of the recent false claims suggesting that a Brother firmware update may have restricted the use of third-party ink cartridges. Please be assured that Brother firmware updates do not block the use of third-party ink in our machines.”

Brother maintains that while they encourage the use of Brother-branded toner for optimal performance, their printers do not intentionally degrade print quality based on the cartridge brand. The company attributes user concerns to its standard troubleshooting process, which includes a “Brother Genuine check”, i.e. a diagnostic step that ensures a printer functions correctly with its own branded consumables.

Not the First Manufacturer to Face Backlash

Brother is not the first printer manufacturer accused of anti-consumer practices related to third-party ink. HP, for example, has implemented “Dynamic Security,” which blocks non-HP cartridges entirely. The company has faced multiple lawsuits over this strategy, with critics accusing it of monopolistic practices. HP, however, defends its approach, citing security risks associated with third-party cartridges, including claims that they could potentially introduce malware.

Similarly, Canon has been criticised for implementing firmware updates that disable scanning functions when ink is low, even when the ink is unrelated to the scanning process. Epson has also faced scrutiny for its use of DRM (Digital Rights Management) technology, which prevents cartridges from functioning after a certain period, even if they remain full.

What This Means for Your Business?

For businesses relying on printers for everyday operations, Brother’s alleged restrictions could pose challenges. Many companies opt for third-party toner to cut costs, often paying significantly less than the high price of OEM cartridges. However, with firmware updates now potentially limiting this option, businesses may find themselves locked into purchasing expensive Brother-branded supplies.

While Brother insists that these changes do not impact third-party ink compatibility, user reports suggest otherwise. Companies using Brother printers may need to consider steps such as disabling automatic firmware updates to prevent potential functionality losses. However, doing so carries security risks, as firmware updates often include patches for vulnerabilities.

The growing controversy surrounding printer manufacturers and third-party ink restrictions could push regulatory bodies to take action. Consumer advocacy groups have already called for stricter oversight of post-sale function removal, arguing that disabling previously available features through software updates constitutes deceptive business practices.

As the debate unfolds, businesses should remain vigilant, assessing how these firmware updates might affect their operations. While Brother’s printers have historically been a more user-friendly option in terms of third-party toner compatibility, recent developments suggest that the industry as a whole is moving toward tighter control over consumable use, thereby raising important questions about fairness, transparency, and consumer choice.

GitLab Threat Intelligence has warned over 3.2 million Chrome users to delete 16 malicious browser extensions that compromise security and expose data to attackers.

The extensions, including ad blockers, screen capture tools, and emoji keyboards, were found injecting harmful code into browsers. GitLab says attackers hijacked these extensions through phishing or by acquiring them from developers, using them to bypass security protections and manipulate content.

Once installed, the extensions connect to a remote server, receive hidden commands, and strip security measures from websites. Although Google has removed them from the Chrome Web Store, the warning to users is to manually uninstall them to stay protected.

Compromised extensions include Blipshot, WAToolkit, Super Dark Mode, and Adblock for Chrome. Experts warn that high download counts and positive reviews do not guarantee safety, as attackers often hijack trusted extensions.

Businesses should restrict unverified extensions, review permissions regularly, and use endpoint security to prevent such threats. Monitoring browser activity can also help detect potential risks early.