News

AI

News about AI: events, facts and developments on tafs uz.

158 stories

Anthropic: surge in vulnerabilities threatens critical infrastructure with a 'tsunami'

Anthropic's deputy chief information security officer Jason Clinton warned that a wave of newly discovered software vulnerabilities could overwhelm defense systems in the coming months, putting power grids and water supplies at risk. He urged isolating sensitive networks and asking boards of directors for help with 'the basics' rather than AI, estimating the period of elevated risk at roughly 18 months.

2 min read

Bloomberg values Anthropic investors' stakes ahead of IPO: from Schmidt to Sequoia

According to Bloomberg, even before a possible IPO of Anthropic at a valuation of up to $2 trillion, its investors' stakes are worth billions: Eric Schmidt owns roughly 1% of the company (about $10 billion), Iconiq holds $17 billion, Menlo Ventures $14 billion, Sequoia about 1% after a $3 billion investment, and Sound Ventures and Spark Capital have also seen multiple returns on their investments.

2 min read

IT professionals on the PHP emergency: “It was endless”

After two critical vulnerabilities were discovered in PHP, the language powering millions of websites, IT teams worldwide rushed to install patches or take servers offline. Satya Nadella called the situation unprecedented, but experts note that rumors of hacker attacks — some of which were not confirmed — amplified the panic.

3 min read

YouGov survey: French children use AI noticeably more often than their parents think

A YouGov survey commissioned by the Berlin-based NGO Interface showed that parents in France underestimate their children's AI use: 40% of children aged 11–12 use general-purpose AI services weekly, while only 26% of parents are aware of it. 42% of parents feel no support on this issue, and 51% believe schools should bear the responsibility. Against the backdrop of the results, the EU's proposed Kids Act is being discussed.

4 min read

UK AI minister: artificial intelligence must not be left to 'techno-feudalists and oligarchs'

AI minister Kanishka Narayan told the BBC that control over AI should belong to society and democracy, not a narrow group of tech elites. He compared the scale of change to the industrial revolution, described the discovery of 500 vulnerabilities in government systems using AI for £13,000, and hinted at an AI announcement in the budget.

3 min read

AI finds two critical PHP vulnerabilities: emergency updates and regulator warnings

Researchers from Calif and Wiz, using AI, discovered two critical vulnerabilities (CVE-2026-109618 and CVE-2026-109619, rated 9.2/10) in the PHP language, affecting versions from 5.4 to 8.5.11. Developers released emergency patches on October 21, and regulators in the US, UK, EU and Australia urged urgent system updates. Cloudflare recorded attacks after details leaked ahead of the official disclosure.

3 min read

AI companies urged UK government to prevent data center protests within 12 months

The Sovereign AI Leadership task force, which includes Nvidia, Microsoft and Cisco, warned the UK government that the country has about 12 months to secure public support for building data centers. The appeal, reported by The Guardian, notes growing protests over energy bills, water use and the environment. Campaigns criticized the idea of financial incentives for local authorities.

3 min read

a16z renames its crypto division: a16z Crypto becomes a16z Frontier

Andreessen Horowitz announced the renaming of its digital assets division a16z Crypto to a16z Frontier. The team, led by Chris Dixon, which has raised nearly $10 billion since 2018, is expanding its focus to new technologies — AI, robotics, and networks connecting people and machines — but is not abandoning cryptocurrencies.

1 min read

British campaign Pull the Plug prepares march against machines in London on December 5

The British campaign Pull the Plug is organizing a protest outside the BBC building and a march through central London on December 5 against the rapid spread of AI. The group demands a moratorium on new data centers, the convening of citizens' assemblies to develop AI regulation, and criticizes the influence of OpenAI, Anthropic, Meta, Microsoft and Amazon on policy.

3 min read

Railway platform urges all users to revoke secret keys after cyberattack

The development platform Railway said that as a result of a cyberattack, attackers may have gained access to encrypted environment variables of some customers, and for some of them, to decrypted content. The company urged all users to revoke and replace API and cloud keys and paused deployments during the investigation. Railway did not disclose the number of affected users or the cause of the incident, suggesting the attack was linked to a campaign against AI providers.

2 min read

PHP Vulnerability: Websites Worldwide Taken Offline as a Precaution

Vulnerabilities that had existed for years have been discovered in the popular programming language PHP, which powers millions of websites. They are difficult to exploit, but the companies Cloudflare and Palo Alto Networks are recording millions of attack attempts. A patch has been released, but it only works on the latest version of PHP, so older websites require an update. Experts describe the situation as 'the cyber equivalent of an unexploded bomb'.

3 min read

France Establishes the Prix Pascal Award for Books Written in Co-Authorship with AI

Amid the scandal surrounding Télesphore Orléane and debates over AI among Goncourt Prize finalists, France has launched the Prix Pascal de Littérature IA, openly encouraging the use of AI in writing books. Participants must submit a manuscript and a technical dossier describing the algorithmic assistance. The 2026 winner will be announced in mid-December; there is no cash prize — help with publication is promised instead.

3 min read