Updated daily503 verified stories
Safety News tracker.
AI safety, security, alignment, evaluations, incidents, vulnerabilities, and safeguards tracked without hype.
- 54stories this week
- 503sources cited
- 503verified timeline entries
Verified sourcing
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
Plain English
What happened, why it matters, and what to watch — without the jargon.
No filler
When the signal is thin, we publish nothing rather than padding the feed.
More stories
9 storiesSecurity
Bill Gates warns AI‑driven cyberattacks will dominate security challenges this year
In an ABC News interview on 11 October 2026, Microsoft co‑founder Bill Gates said the biggest AI‑related threat in the coming year will be malicious cyber‑attacks, arguing that current security measures are inadequate and that industry self‑regulation is insufficient.sudouest.frSecurity
Japan issues nationwide cyber warning as AI‑assisted attacks surge
Japan’s government warned ministries, local authorities and private firms to tighten cyber defences after a sharp rise in data breaches and AI‑enabled attacks, citing incidents at Daiwa Securities, SoftBank, Lawson and Times Car.the420.inSecurity
Bill Gates warns AI‑driven cyberattacks could surge within a year
In an ABC News interview, Microsoft co‑founder Bill Gates said malicious actors will likely use artificial intelligence to multiply cyberattacks, making them the top security challenge in the coming year.portafolio.coSecurity
Booz Allen tests show AI agents can breach industrial control systems
Booz Allen Hamilton's OT lab found that frontier AI models can autonomously compromise SCADA systems and robotic arms in a simulated manufacturing environment, exposing significant gaps in critical infrastructure security.news.lavx.huIndustry
AI safety evaluators gain prominence as labs face internal and external pressure
As major AI labs like Anthropic and OpenAI move to embed third-party safety evaluators, the industry faces unresolved questions regarding funding, independence, and the scope of access for these auditors.cnbc.comEnterprise
University of Maine System signs two‑year ChatGPT Edu contract with OpenAI
The University of Maine System has entered a two‑year pilot with OpenAI for ChatGPT Edu, costing about $22 per user per year and promising data‑privacy safeguards for students and staff.mainepublic.orgSecurity
Researchers use explainable AI to surgically break LLM safety filters
A new study published in Neural Computing and Applications demonstrates a white-box attack technique called XBreaking that uses explainable AI to identify and disable safety alignment in open-source large language models with high success rates.bioengineer.orgProduct
ONLYOFFICE 10.0 adds context‑aware AI tools and modular apps to its open‑source suite
ONLYOFFICE’s autumn release 10.0 embeds AI assistants directly in its document, spreadsheet, presentation and PDF editors, adds four free modular apps and patches 35 security flaws, expanding the open‑source office platform’s AI and security capabilities.bastillepost.comSecurity
Hackers use Google ad and Bing redirect chain to deliver fake Claude installer
Push Security uncovered a novel malvertising campaign that starts with a Google‑sponsored ad for “claude mac”, routes through a Bing click‑tracking redirect and a compromised retailer site, and ends on a convincing fake Claude download page that installs malware.pushsecurity.com
One useful briefing each week
Keep up with AI without living in the feed.
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Reach people who are learning AI
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Post an AI jobSubmit an AI tool