AI Daily Brief: 26 May 2026

26 May 2026

Three distinct power struggles dominate today's AI landscape. The UK Chancellor is pushing back against her own procurement machinery to put British AI firms first. The Vatican is challenging Silicon Valley's grip on AI development in a 42,000-word papal document. And Palantir is fighting City Hall - literally - after losing a £50m Met Police contract.

Reeves orders ministers to 'buy British' for AI and defence contracts

Chancellor Rachel Reeves has written to every cabinet minister in charge of a spending department, instructing them to prioritise British companies when awarding contracts in four industries: shipbuilding, steel-making, energy, and artificial intelligence. The Treasury and Cabinet Office will now monitor billions of pounds worth of contracts and can override departmental decisions if ministers do not comply.

The letter, seen by the Guardian, cites specific examples that have frustrated Reeves - a £200m naval support vessel contract awarded to Dutch firm Damen, a £9m refit for the research ship Sir David Attenborough signed with a Danish yard, and ongoing uncertainty over a £1.9bn Faslane nuclear submarine facility upgrade that could go abroad. On AI specifically, the directive signals that the government wants British AI companies to be the default for public sector contracts rather than US or other foreign suppliers.

For UK AI firms, this represents a meaningful structural shift. Government contracts have historically gone through competitive tender regardless of national origin. The new direction - backed by Treasury authority to override ministers - gives domestic AI companies a structural advantage in public sector sales for the first time.

Our take: This is a significant signal for the UK AI sector. 'Buy British' guidance from the Chancellor, with override powers attached, is far more than a policy preference - it is a procurement doctrine. UK AI companies that have struggled against better-resourced US competitors for government contracts should be paying close attention. The most commercially interesting question is what counts as a 'British' AI firm in complex multi-vendor arrangements. Expect that definition to be contested.

Pope Leo XIV issues 42,300-word encyclical demanding AI arms control

Pope Leo XIV has released 'Magnifica Humanitas' - a 42,300-word encyclical calling for the most rigorous ethical constraints on AI in warfare, and warning that some autonomous weapons systems are now 'practically beyond any human reach' to control. Leo, the first American pope, presented the document himself at the Vatican. Christopher Olah, co-founder of Anthropic, was present alongside the pope at the formal launch - one of the more unexpected images of 2026.

The encyclical warns that power over AI and digital infrastructure 'does not rest with states but with major economic and technological actors', and that when concentrated 'in the hands of the few' it risks becoming 'opaque and evading public oversight'. Leo called for AI data ownership not to be left solely in private hands, urged a cooling of competition between AI companies, and described the technology's spread as part of a broader 'culture of power' risking new forms of dependency and inequality.

Olah, speaking alongside the pope, said there was 'a real possibility' that AI would displace human labour 'at very large scale', and acknowledged that AI companies operate inside commercial pressures that can sometimes conflict with doing the right thing for society - making outside scrutiny essential. The encyclical is one of the highest forms of papal teaching and is addressed to the Catholic church's 1.4 billion members.

Our take: Whatever your view of the Vatican's institutional authority, a 42,300-word document read by 1.4 billion people calling for AI arms control and challenging tech concentration is a geopolitical event, not a religious curiosity. UK business leaders navigating AI governance conversations with boards and regulators may find the encyclical's language - particularly on concentration of power and worker displacement - appearing in places they did not expect. The presence of an Anthropic co-founder at the launch suggests the AI industry is actively cultivating this relationship.

Palantir accuses Sadiq Khan of 'putting politics above public safety' after Met contract blocked

Palantir has hit back at London Mayor Sadiq Khan after he blocked the company's £50m contract with the Metropolitan Police. Khan rejected the deal, citing what he described as a 'clear and serious breach' of procurement rules. Palantir's response was direct: the company accused the mayor of prioritising political considerations over the safety of Londoners.

Since our reporting on 25 May, the new development is Palantir's formal public counter. The company argues that AI-assisted intelligence analysis would have made the capital safer, and that blocking the contract on procurement process grounds - rather than on the merits of the technology - represents a politically motivated decision.

The row highlights a wider tension in UK public sector AI adoption: procurement rules designed for traditional contracts are increasingly being applied - or used as grounds for blocking - AI and data analytics agreements where vendor selection involves considerations that standard tendering frameworks were not built to handle. The Met case is likely to be cited in future debates about how UK local and national government structures AI procurement.

Our take: Palantir's public counter-attack is a calculated move. By framing Khan's decision as 'politics above safety', they are shifting the debate from procurement compliance to public consequences. Whether that framing lands will depend heavily on what the 'clear and serious breach' actually involved - and that detail has not yet been made public. For other UK public bodies considering AI contracts with US firms, this episode is a reminder that procurement process is not a technicality. It is the battleground.

'AI washing': UK firms are rebranding ordinary automation as artificial intelligence

PR executives at UK agencies are reporting pressure from clients to rebrand standard business software and automation tools as 'artificial intelligence' in order to attract investment, talent and customers. The Guardian reports companies across multiple sectors are asking communications teams to present rule-based systems, basic data analysis pipelines, and automated scheduling tools as AI products or AI-driven services.

The phenomenon mirrors the 'greenwashing' problem in sustainability communications, where firms overstate their credentials to benefit from a popular trend. Industry observers say the pressure has accelerated sharply since 2024, as investor scrutiny of AI capability became a standard part of due diligence and job postings that include AI terms attract significantly more applicants than comparable roles without them.

For UK businesses evaluating supplier claims, the practical implication is clear: the label 'AI-powered' or 'AI-driven' has become almost meaningless without independent verification. Asking vendors to specify which model, what training data, and what measurable outcomes the AI component actually produces is now a basic procurement hygiene step rather than an optional technical deep-dive.

Our take: The AI washing story is the quiet companion to every headline about adoption rates. When research shows 54% of UK SMEs say they 'use AI', a meaningful proportion of that figure reflects software with an AI badge rather than genuine machine learning capability. For real AI companies - and for buyers trying to evaluate them - this noise is commercially damaging. Expect the FCA, CMA and sector regulators to begin treating AI capability claims with the same scepticism they now apply to sustainability claims.

Quick Hits

• HSBC CEO told staff to stop fighting AI as the bank accelerated its workforce change programme, continuing a trend of banking leaders becoming more direct about AI-driven job changes.
• Spotify and Universal Music Group reached a landmark licensing deal allowing Premium subscribers to create AI covers and remixes of songs - the first major licensed AI remix tool for consumers at scale.
• 342 tech companies have announced layoffs in 2026 so far, affecting nearly 144,000 workers - a faster pace than the whole of 2025 - as firms restructure around AI automation rather than headcount growth.
• The White House reversed its position on requiring safety reviews of new AI models in Trump's executive order after sustained tech industry lobbying, clearing the path for faster model releases without federal pre-clearance.

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.