AI Daily Brief: 31 May 2026

31 May 2026

The AI industry's money is moving at a scale that makes 2021 look like a dry run. Anthropic has eclipsed OpenAI in valuation, SoftBank is committing $87 billion to French data centres, and Samsung chip workers are walking away with £310,000 bonuses. Meanwhile, UK banks still cannot access the AI model the Bank of England says could rewrite cyber risk - and a major new TUC-backed report warns British workers are already losing jobs without any say in how AI is being deployed.

UK Banks Still Blocked from Anthropic's Mythos AI After Six Weeks

British banks still cannot access Anthropic's Mythos AI model to test their systems against cyber threats, Bank of England Governor Andrew Bailey said on Friday - six weeks after he first raised the alarm. Speaking at a central banking conference in Reykjavik, Bailey said Anthropic was willing to share the model on a trial basis but the process had been caught up in US political delays.

"It hasn't happened yet and I think this has been somewhat caught up in the process with the US administration," Bailey told Bloomberg TV. The BoE governor has previously said Mythos could "crack the whole cyber risk world open." He called for a coordinated international response, warning that cyber spillovers are "so big that we can't just have a single sort of national approach."

The holdup matters directly for UK financial services firms trying to assess their own vulnerabilities. Without access to the model, banks cannot run defensive red-teaming exercises - the kind of proactive security testing regulators increasingly expect. President Trump reportedly postponed signing a broader AI executive order last week that would have created a framework for government-AI company engagement ahead of major model releases.

Our take: Bailey is right that this cannot be solved country by country, but the practical message for UK financial services leaders is stark: you cannot wait for a geopolitical resolution before stress-testing your defences against AI-capable threats. If Mythos-level capabilities are coming to market in weeks - as Bloomberg reports Anthropic plans - the window for defensive preparation is closing fast.

Anthropic Overtakes OpenAI with $965bn Valuation After $65bn Raise

Anthropic has raised $65bn in a Series H funding round at a post-money valuation of $965bn, making it the world's most valuable AI startup and leapfrogging OpenAI. The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital - and marks a sharp reversal of fortunes for a company once seen as the principled but commercially modest counterpart to its larger rival.

The growth has been driven primarily by enterprise adoption, particularly following Anthropic's release of powerful coding tools late last year. "Claude is increasingly indispensable to our growing global community of customers," said CFO Krishna Rao. The valuation arrives as Anthropic maintains an active legal dispute with the Pentagon over its refusal to allow Claude to be used in lethal autonomous weapons systems or domestic mass surveillance - a stance that appears to be helping rather than hurting its commercial standing.

The funding round will reshape the IPO landscape for the rest of 2026. Both Anthropic and OpenAI have been expected to list publicly this year. With Anthropic now valued well above its rival, the competitive dynamics of the industry have shifted in ways that will be felt by every enterprise choosing an AI platform.

Our take: Anthropic's rise to become the most valuable AI firm is a direct result of a clear strategic bet: go deep on enterprise and safety rather than consumer apps. The $965bn valuation signals that institutional capital believes the enterprise AI market is more durable than the consumer race. For UK businesses evaluating AI suppliers, Anthropic's positioning - including its willingness to refuse military misuse contracts - now carries real commercial weight alongside the technical capabilities.

SoftBank Commits up to €75bn for AI Data Centres Across France

SoftBank Group announced on Friday it will invest up to €75bn ($87bn) to build 5 gigawatts of AI data centre capacity in France. The initial phase - €45bn over five years - focuses on facilities in Dunkirk, Bosquel, and Bouchain, with additional sites planned across France in later phases. The announcement came alongside a meeting between SoftBank founder Masayoshi Son and French President Emmanuel Macron.

France has moved aggressively to attract AI infrastructure investment since the global success of Mistral AI, positioning itself as the leading European destination for large-scale AI compute. The SoftBank commitment follows a series of major announcements at the Choose France investment summit held earlier in May. The combined planned investment from the summit now runs to hundreds of billions of euros.

The scale of the SoftBank deal - comparable to entire national AI strategies - underlines the infrastructure race now underway among leading AI nations. For the UK, which has its own AI infrastructure ambitions but has struggled to match the pace of US or French commitments, this announcement adds pressure on the government to accelerate data centre planning approvals and energy capacity decisions.

Our take: France is winning the European AI infrastructure race, and the SoftBank deal is the clearest proof yet. Compute capacity is becoming a strategic national asset. The UK government's AI Opportunities Action Plan acknowledged this - but acknowledgement and execution are different things. Every month of planning delays or grid capacity constraints is a month where the competitive gap with France widens.

4% of UK Workers Say They Have Already Lost Jobs to AI, TUC-Backed Report Finds

A new report from the Institute for Public Policy Research (IPPR), backed by the TUC, has found that 4% of British workers believe they have already lost a job because of AI - while 21% say the technology has made their working life worse. Only 20% said AI had improved their experience at work. The report calls for a statutory duty on employers to consult workers before adopting AI, alongside a portable "worker support levy" to fund training, union membership, and reskilling.

The IPPR distinguishes between three ways AI affects work: augmentation (complements human labour), degradation (undermines job quality through monitoring and algorithmic management), and displacement (replaces workers altogether). "The question is not whether AI will disrupt working life, but who will have the power to shape that disruption," the authors write. TUC General Secretary Paul Nowak invoked the Industrial Revolution as a warning: "50 years of wage stagnation while profits soared. It took the difficult birth of the labour movement to tip technological gains towards workers' interests."

The government has made clear it is enthusiastic about AI adoption - Rachel Reeves has called it "the defining technology of our era" - but the IPPR report argues that existing employment law was not designed for algorithmic management or automated redundancy decisions. A dedicated AI workplace framework, separate from the broader Employment Rights Bill, is what the thinktank is calling for.

Our take: The 4% figure is probably an undercount - workers may not always attribute job loss directly to AI when the visible mechanism is a restructuring or an 'efficiency' programme. The IPPR's recommendation for statutory consultation rights is the most actionable part of this report. UK business leaders adopting AI without any formal staff engagement process are not just taking a reputational risk. They may soon be taking a legal one.

Pope Leo's AI Encyclical Draws Global Responses Six Days After Vatican Launch

Pope Leo XIV's 42,300-word encyclical, released on 25 May, continues to generate debate across business and technology circles this weekend. The document - among the most authoritative forms of papal teaching - calls for AI to be "disarmed" and urges governments, corporations, and individuals to subject the technology to the "most rigorous" ethical constraints. Leo draws a clear distinction between human and machine: "So-called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships."

The encyclical was presented at the Vatican with testimony from Christopher Olah, co-founder of Anthropic - the same week Anthropic announced its $965bn valuation. That juxtaposition has prompted pointed commentary on whether safety-focused AI firms can simultaneously pursue near-trillion-dollar ambitions and principled restraint. Guardian readers in the US described fears about "unregulated AI" threatening workers, privacy, and "human life itself." A Guardian columnist wrote that the encyclical is "an important warning" while Silicon Valley's dismissal of it is "wrong."

For UK business leaders, the encyclical is unlikely to change procurement decisions directly. But it crystallises a governance question that will: at what point does deploying AI without meaningful human oversight become a reputational, regulatory, or ethical liability? The IPPR report published this week and the UK government's own AI governance consultations both point in the same direction - the Pope's document adds significant moral and political weight.

Our take: The significance of this encyclical is political, not theological. When the world's most prominent moral authority devotes 42,000 words to AI governance, it shifts what is considered a reasonable position on regulation. Every government and board that has been deferring an AI ethics policy now faces a harder question to dismiss. The challenge for UK businesses is not to answer the Pope's concerns directly, but to ask honestly whether their own AI governance would survive public scrutiny.

Samsung Chip Workers to Receive £310,000 Bonuses Each in AI Profit-Sharing Deal

Unionised workers at Samsung Electronics' memory chip division have ratified a landmark profit-sharing agreement that will see 62,616 employees receive average bonuses of approximately £310,000 each - drawn directly from surging demand for high-bandwidth memory chips used in AI systems. Samsung will distribute roughly 40 trillion won (around $26.6bn) to chip division staff. The deal averts a potential strike and arrives as SK Hynix and Micron both joined the $1 trillion market cap club on the back of AI-driven chip demand.

The agreement represents one of the most substantial direct transfers of AI-generated value to frontline workers anywhere in the global economy. Its significance lies in two features: the scale of individual payouts (£310,000 on average is genuinely life-changing), and the mechanism - collective bargaining with a recognised union, not a discretionary bonus scheme controlled by management. Samsung did not choose to share this wealth voluntarily; workers negotiated for it.

The context matters for UK businesses. AI adoption is generating enormous wealth creation at the compute layer, but that wealth has so far been concentrated heavily in equity holders and senior executives. Samsung's deal shows that organised labour, in the right sector with the right leverage, can capture a meaningful share of AI upside. Paired with the IPPR report published this week, it frames a question that UK employers will find harder to avoid: what is your equivalent of this deal for the staff whose work is being transformed by AI?

Our take: The Samsung deal will be cited in UK workers' rights and AI debates for years. It is proof that AI-driven profit can be shared at scale with frontline workers when bargaining power exists. For UK employers, this is both a benchmark and a warning: the gap between how AI wealth is distributed in manufacturing and how it is distributed in British services is already a political flashpoint - and it is about to become a policy one.

Quick Hits

• Anthropic released Claude Opus 4.8 on 28 May with improved reliability, coding, and tool use - and plans to release Mythos-level capabilities to the public in coming weeks.
• Google launched Gemini 3.5 Flash on 19 May, a faster lower-cost model designed for latency-sensitive applications including chatbots and real-time agents.
• Gareth Edwards, director of Rogue One, told Amazon's AI on the Lot event that AI is "like a billionaire on acid" and will be better than CGI for filmmaking.
• French multinational Schneider Electric used AI in manufacturing to increase worker productivity rather than cut headcount, a model the New York Times examined as an alternative to job displacement.
• Anthropic opened a Milan office on 27 May to support Italian enterprise customers, researchers, and developers as it expands European operations.

Frequently Asked Questions

How often is the AI Daily Brief published?

Every morning at 7:30am UK time, covering the previous 24 hours of AI news from over 30 sources.

How are stories selected?

UK-relevant stories are prioritised first, then by business impact and practical implications for UK organisations adopting AI.

Why should business leaders follow AI news?

AI is moving faster than any technology in history. Staying informed is essential for making smart decisions about AI investment, adoption, and governance.