Robert Booth

Wynd Kaufmyn, a 69-year-old Berkeley activist, became the first person believed to be jailed for protesting artificial intelligence after being convicted over a 2025 StopAI sit-in that blocked OpenAI’s San Francisco headquarters. She received a one-week sentence after rejecting a necessity defense, while San Francisco prosecutors argued that the protest endangered public safety. Kaufmyn says AI companies knowingly risk catastrophic harm and urges OpenAI, Anthropic and Meta executives to “regain your humanity” and support a global ban on the race to superintelligence. AI-safety experts, researchers and Senator Bernie Sanders have voiced related concerns about inadequate safeguards and uncontrolled development, although the technology industry remains driven by commercial and geopolitical competition. StopAI’s campaign has also faced controversy over alleged disagreements about the use of violent tactics.

A Guardian undercover investigation alleges that CT Group, a lobbying firm co-founded by former Conservative strategist Lynton Crosby, offered payment for favourable client coverage in British newspapers and had previously paid journalists to place apparently independent stories. The allegations raise concerns about covert corporate influence on public opinion and the integrity of journalism. CT Group denies engaging or paying the undercover reporter and says the events were deliberately mischaracterised.

An undercover Guardian investigation reports that CT Group, a lobbying firm co-founded by Lynton Crosby, offered a freelance journalist payments for newspaper stories favorable to property-sector clients. Staff allegedly proposed £100 for each successfully published article, a £500 roundtable fee and editorial input, while advising that the firm’s involvement need not be disclosed to editors. The proposed coverage would criticize the Building Safety Regulator and promote the interests of Live-in Guardians. CT Group denied paying or engaging the reporter, said the allegations were deliberately mischaracterized, and maintained that it expects journalists to follow professional standards. The revelations come amid wider scrutiny of lobbying transparency and conflicts of interest in British public affairs.

A global malfunction in Amazon Web Services’ billing system generated wildly incorrect estimated charges, with some customers receiving apparent bills in the billions and trillions of dollars. Users in the UK, India and the US reported shock and distress as routine low‑cost services appeared to incur astronomical fees. AWS acknowledged a pricing computation error, disabled the estimation system and began recalculating data, apologising for the confusion while incorrect figures continued to appear for several hours.

Ofcom has opened a formal investigation into TikTok over concerns that its age verification systems fail to reliably identify children, potentially exposing them to harmful content such as self‑harm, disordered eating and pornography. The probe follows the rollout of the UK’s Online Safety Act and comes as the government prepares a social media ban for under‑16s. Ofcom expressed doubts about age‑inference tools used across multiple platforms and urged companies to adopt more effective verification methods. Meta announced new alerts for parents when children discuss self‑harm with AI chatbots, while search engines Google and Bing are working with Ofcom to address the ease of finding pornography sites lacking age checks. The investigation could lead to significant fines or access restrictions if TikTok is found non‑compliant.

The UK government plans to encourage 16- and 17-year-olds to follow a default midnight-to-6am social media curfew as part of broader efforts to limit online harms and improve sleep. The measure extends earlier restrictions for under-16s and targets platforms such as Snapchat, TikTok and Instagram, though teens can override the curfew. Campaigners criticized the policy as insufficient, while officials emphasized its role in supporting wellbeing. Research commissioned by the government found that curfews were the most manageable restrictions and improved sleep quality, though some young people reported feeling disconnected from peers.

Two parliamentary committees and more than 100 NHS data experts are urging the government to abandon Palantir’s £330m NHS data platform contract, citing inadequate privacy protections, mistrust among patients and clinicians, and alternative options. Critics argue Palantir’s military work and ethical concerns undermine public confidence, while supporters claim the platform improves hospital efficiency and patient outcomes. Labour is being pressed to use a 2027 break clause to end the deal as unions and MPs call for the NHS to retain control of sensitive health data, while the government continues to defend the platform’s performance gains.

UK information commissioner John Edwards, who resigned after an investigation upheld multiple allegations of sexual harassment and bullying, is preparing legal action against a woman who reported his behaviour. Science and technology secretary Liz Kendall condemned the move and announced an independent review into the ICO’s culture and governance, alongside plans for a majority‑women board. Kendall also outlined forthcoming social media curfews for 16‑ and 17‑year‑olds and highlighted broader concerns about disinformation, prompting an urgent Ofcom capability review.

A proposed Scottish National party moratorium on new datacentres could jeopardize the UK’s AI infrastructure plans, particularly projects relying on Scotland’s renewable energy capacity. Local leaders warn that existing proposals amount to overdevelopment, while revelations emerged about misrepresented feasibility claims for major sites such as the Lanarkshire AI growth zone. The move comes amid broader uncertainty surrounding UK technology policy as Andy Burnham prepares to enter Downing Street. Parliamentary scrutiny has highlighted ineffective investment vetting, overhyped AI growth zones, and growing risks to UK sovereignty following US restrictions on access to advanced AI tools. A cross-party committee has urged the incoming government to create a clear international technology strategy as details show that nearly half of the firms supported by the £500m Sovereign AI Fund are ultimately US-controlled.

Researchers from Oxford and Potsdam found that major AI drafting tools alter the meaning of user messages on sensitive issues such as abortion, religion, climate change and gender roles, introducing political biases even when instructed to preserve intent. Systems from Meta, Google, Alibaba and Mistral tended to shift posts in a liberal direction, while Grok on X leaned toward right‑leaning interpretations. The study warns that small AI‑introduced nudges could accumulate across platforms, potentially reshaping public opinion and falling outside existing regulatory frameworks such as the EU AI Act. Experts caution that AI tools may distort human communication by subtly modifying users’ intended messages.

AI-generated summaries on Tripadvisor hotel pages were found to downplay serious guest complaints, including food poisoning, poor hygiene, and sexual harassment, according to an investigation by Which?. Examples included describing a hotel facing a mass food-poisoning lawsuit as spotless and calling service at a hotel with harassment reports friendly. Tripadvisor said it is monitoring and refining its AI tools and that summaries are meant to complement, not replace, detailed reviews. Experts noted that AI summarization often dilutes strong criticism. The findings follow broader concerns about AI-generated summaries across industries, including recent issues with Google’s health summaries.

US export restrictions on Anthropic’s advanced AI models Fable 5 and Mythos 5 were lifted after government review concluded enhanced safeguards were sufficient. The temporary ban, imposed over fears the systems could assist in sophisticated cyber‑attacks, had limited access for foreign nationals and delayed broader deployment. The decision followed collaboration between Anthropic and the US Department of Commerce to strengthen security and create standards for future model releases. Intelligence agencies from Five Eyes countries warned of rapidly escalating cyber risks associated with frontier AI. Industry figures, including Sam Altman, criticized government control over customer vetting as OpenAI delayed wider release of GPT‑5.6 at federal request.

Concerns are raised that the Financial Conduct Authority’s partnership with Palantir could expose sensitive UK financial and personal data to the US government under American disclosure laws, particularly during a Trump administration. MPs and digital rights advocates warn that the US Cloud Act and related legislation could compel access despite FCA assurances that data will remain encrypted and under its control. Palantir disputes the risks, citing legal and technical safeguards, while critics argue that foreign-controlled data processing still poses sovereignty and privacy vulnerabilities.

The UK government is preparing new restrictions on children’s access to social media, with measures such as age limits and controls on addictive design features expected by the end of the year. Following extensive consultation responses and lobbying from technology companies, Labour aims to act quickly amid internal and external pressure. Advocates and bereaved families remain divided over an Australia-style ban for under‑16s, with some warning of safety risks and others urging immediate action. Ministers continue to meet stakeholders, including big tech firms, as they finalize proposals that could require platforms to demonstrate child‑safe design or face being blocked for young users.

Wes Streeting urges treating social media like tobacco and calls for banning under-16s from certain platforms as the UK government nears a decision on age limits. He argues large tech companies are avoiding regulation and warns of health risks linked to online content. Medical bodies report rising harms among children and advocate tighter controls, while children’s organisations say age limits alone are insufficient, urging bans on personalised services for under‑13s. Bereaved parents present differing views on the proposed ban as ministers weigh options including feature restrictions, curfews and algorithm limits.

Palantir condemned Sadiq Khan’s decision to block a £50m Metropolitan police contract, accusing the London mayor of politicising procurement and undermining public safety. Khan’s office cited a serious breach of procurement rules, while critics questioned Palantir’s values and involvement with foreign governments and militaries. The move exposed divisions within Labour, with some MPs supporting Khan’s stance and others praising Palantir’s technological capabilities. Scotland Yard warned that rejecting the contract could affect policing capacity, while concerns grew over expanding reliance on foreign AI providers.

London’s mayor blocked a £50m contract between the Metropolitan police and Palantir due to breaches in procurement rules, citing risks of vendor lock-in and concerns about the company’s ethical record. Scotland Yard warned the decision could hinder its ability to modernise and maintain officer numbers amid a budget shortfall. Critics highlighted Palantir’s history of expanding its influence through small or free initial contracts, while supporters of the block argued it ensures proper scrutiny of major public spending. City Hall will work with the Met on a new procurement plan, and Palantir remains able to bid for future contracts.

Growing reliance on artificial intelligence in UK policing has sparked controversy over the Metropolitan police’s proposed £50m contract with Palantir, blocked by London mayor Sadiq Khan over procurement and ethical concerns. Scotland Yard argues AI could offset staffing shortfalls, while government ministers push for rapid adoption of AI across public services. Palantir’s ties to contentious US government actions and its leaders’ provocative statements have fueled opposition among politicians, the public and police unions. Critics warn of overreliance on a single US tech firm, while experts note that British alternatives remain less developed due to lack of investment.

Anthropic co-founder Jack Clark predicted rapid advances in AI, including a Nobel-level scientific breakthrough within a year, AI-run companies generating significant revenue, and bipedal robots assisting workers. He warned of existential risks, urging preparation for increasingly powerful systems amid global competition. At an Oxford University lecture, he compared society’s unpreparedness to past pandemic failures, while experts cautioned that growing reliance on AI could lead to cognitive decline. Concerns were raised about concentrated control of frontier AI models and the broader societal impact of accelerating technological change.

Google DeepMind agreed to formal talks with UK unions after London-based staff voted to pursue unionisation amid rising concerns over the use of its AI systems by the US and Israeli governments. The company declined voluntary union recognition but will participate in Acas-facilitated discussions that could lead to a formal ballot. Worker petitions highlight fears about military and intelligence applications following Google’s rollback of restrictions on harmful weapons and surveillance technologies. A dismissed AI researcher is also challenging the company over alleged retaliation for protesting its work for Israel. Unions view the talks as a step toward addressing workplace concerns and scrutiny of government-linked AI contracts.

Child safety advocates including Jonathan Haidt and several organizations filed a complaint urging the US Federal Trade Commission to investigate Roblox, alleging the platform’s design and business model expose children to harm and prioritize engagement and profit over safety. The dossier highlights risks in chat features, monetization practices, and exposure to inappropriate content, while noting Roblox’s rapid growth and reliance on virtual currency sales. The move comes amid heightened scrutiny of major tech platforms, recent legal rulings against Meta and YouTube, and legislative efforts to strengthen protections for minors. Roblox disputes the claims, stating its safeguards and policies are designed to maintain a safe environment and that purchases are optional.

A former Google DeepMind engineer has filed an employment tribunal claim alleging unfair dismissal after protesting the company’s work with the Israeli government, including distributing flyers and urging colleagues to unionize. Google denies firing him over his views, asserting adherence to policies supporting open discourse. Staff concerns reportedly intensified after changes to Google’s AI principles and the expansion of military and intelligence-related AI work. Growing worker resistance has also emerged over AI’s defense applications, including protests against Google and Amazon’s cloud deal with Israel and a call to bar classified U.S. government AI work. The case is supported by the tech justice group Foxglove amid broader public unease about the societal impacts of AI.

Ofcom will strengthen its codes of practice to require technology platforms to detect and block intimate image abuse, including AI‑generated deepfakes, amid rising concerns about harmful content targeting women and girls. The move follows pressure from campaigners and a threatened legal challenge from End Violence Against Women, who argue the regulator must take stronger action. Ministers, including Keir Starmer and Liz Kendall, have called for swift removal of non‑consensual intimate images and stricter enforcement under the Online Safety Act. Campaigners welcome the updated guidance but say binding requirements are needed to stop abusive content and shut down offending forums.