Journalism begins where hype ends

,,

The American people do not want to be tracked and recorded by AI"

—Senator Bernie Sanders

OpenAI Pauses Frontier RL Training for Models Including Astra Over Safety Risks

OpenAI CEO Sam Altman and the company’s Astra model branding, as OpenAI pauses frontier reinforcement learning training over safety and cybersecurity concerns. Source Wiki Commons

OpenAI CEO Sam Altman announced the company is temporarily pausing some frontier reinforcement learning (RL) training, including work on its upcoming Astra model, to ensure alignment, security, and monitoring standards keep pace with rapidly advancing capabilities. The move follows a security incident involving Hugging Face and preliminary findings that Astra may reach “critical” cybersecurity thresholds.

What are the “Most Urgent AI Risks” according to 272 Experts?

AI risks are emerging across vulnerable sectors including finance, national security and information.

A 2026 Delphi study identifies the AI risks considered most severe over the next five years and finds information, finance, and national security among the sectors most vulnerable to AI risks

“Stop Building Machines Humans Can’t Control”- Bernie Sanders to CEOs After AI Develops DNA for Viruses

Senator Bernie Sanders warns that AI systems capable of designing novel viral genomes could enable new bioweapons, urging OpenAI, Anthropic and Meta to pause development before humans permanently lose control.

Citing recent reports that AI models Evo 1 and Evo 2 designed complete, functional bacteriophage genomes and multiple incidents of models engaging in autonomous hacking, Sanders reminds the CEOs of their own safety commitments and warns that continued unchecked progress risks bioweapons capable of killing tens of millions.

‘We Are Living in Don’t Look Up’: Sanders Warns Senate on Unchecked AI

Bernie Sanders AI Warning: Senator Invokes 'Don't Look Up

In a Senate address on 23 July, Bernie Sanders compared artificial intelligence to the extinction-level comet in Don’t Look Up, warning that elected representatives are not paying attention and that the consequences for workers could be catastrophic.

China’s AI Safety Focus Shifts From ‘What AI Says’ to ‘What AI Does’: Report

China is shifting its AI safety strategy beyond regulating AI-generated content toward governing increasingly autonomous AI systems, according to the Concordia AI report.

According to a new Concordia AI report, China is shifting its AI safety focus beyond AI-generated content toward governing increasingly autonomous AI systems and the risks they pose to cybersecurity, society, and critical sectors.

DeepMind and Isomorphic Labs Pitch “Bioresilience” Plan to Keep AI Out of Bioweapon Hands

DeepMind's AlphaFold protein-structure prediction anchors the "response" pillar of the bioresilience plan, which the companies say can speed up vaccine and treatment development.

Google DeepMind and Isomorphic Labs have set out a joint “bioresilience” approach aimed at stopping threat actors from misusing their AI models to develop biological weapons.

Anthropic CEO Dario Contributes $1 Million to ‘Public First’ PAC That Supports Greater AI Regulations

Dario Amodei has contributed $1 million to Public First PAC.

The individual contribution by Dario comes in the wake of Anthropic advocating for a globally coordinated pause on Frontier AI research by AI companies.

Demis Hassabis Proposes Frontier AI Standards Body, Picks US as Host Country

Demis Hassabis, CEO and co-founder of Google Deepmind has called for an international Frontier AI Standards Body to be set up in US for terming AI models Frontier Class.

The path to Artificial General Intelligence (AGI) is challenging but what’s even more challenging is a global consensus on what and who qualifies as “Frontier Class” and who gets to decide that.

“Race to Build Smarter-Than-Human AI Extremely Unpopular”: Hundreds Protest in San Francisco

Protestors from a 13-organisation coalition march through San Francisco on July 11, demanding a halt to frontier AI development.

Hundreds marched to the San Francisco offices of Google, OpenAI and Anthropic on July 11, demanding a moratorium on frontier AI development. In an exclusive conversation with AI FrontPage, Evitable CEO David Krueger said the race to build smarter-than-human AI is “extremely unpopular” and proceeding despite active public resistance.

Switzerland to Hold July Curtain-Raiser Ahead of 2027 Geneva AI Summit

Switzerland will host the 2027 Geneva AI Summit, with a curtain-raiser planned for July 2026.

The event will announce the theme and further details of the 2027 Geneva AI Summit, which will emphasise international law and fundamental rights in the development and deployment of AI systems.