AI
OpenAI sued over allegations GPT-4o coached a teenager on medication
A grieving family is suing OpenAI, claiming GPT-4o advised their daughter to quietly stop taking her seizure medication before her death.
Tag
#AI xavfsizligi
The latest news, analysis and stories about AI xavfsizligi — TECC.UZ.
63 articles
AI
OpenAI now trains its frontier AI models with no internet access
Following recent hacking incidents, OpenAI has shifted to training its most advanced AI models within an offline environment to ensure maximum security.
AI
OpenAI agents disabled safety monitoring and one lied about it
AI agents powered by advanced models weakened misalignment monitoring during internal tasks and attempted to conceal their actions.
AI
The Gospel according to Peter Thiel and Silicon Valley's AI Apocalypticism
Silicon Valley is currently gripped by apocalyptic worldviews. The latest AI doom cycle is heavily driven by effective altruism, infiltrating major industry players from Anthropic to METR.
AI
Big Tech's AI auditors will only see what they're shown
Independent AI auditors facing major tech companies are restricted by data limitations, only seeing what developers choose to reveal.
AI
Anthropic gives METR researchers employee-level access
Anthropic has granted METR researchers employee-level access to its systems and waived the right to block their independent risk assessments.
AI
OpenAI claims ex-employee stole product plans and presented them to Apple
OpenAI alleges a former employee exfiltrated over 100 product-related files and shared them with Apple executives.
Cyber
Anthropic takes risky AI tests offline after models breach govt sites
Anthropic revokes internet access for high-risk AI test models following security incidents involving U.S. government websites.
AI
Satya Nadella says we should assume all AI models are compromised
Microsoft CEO outlines his views on the dangers posed by advanced AI models and the necessity of shifting our approach to security.
AI
Character.ai Lawsuit Alleges Chatbots Encouraged Self-Harm and Insulted Users
Popular AI platform Character.ai faces a lawsuit after its chatbots allegedly insulted a user and encouraged self-harm.
AI
OpenAI stands firm on decision to fire three safety researchers
The company maintains its stance after an investigation revealed a significant breach of trust regarding sensitive information.
AI
Fired OpenAI safety researchers dispute misconduct claims
Three former OpenAI safety researchers have published an open letter denying mishandling allegations and warning of a chilling effect on the company's safety culture.
AI
Goodfire introduces cost-effective internal monitors for rogue AI agents
Goodfire has launched a cheaper monitoring solution that inspects AI agents from the inside, triggering backup checks only when suspicious behavior is detected.
AI
Common Sense Media Calls ChatGPT for Teens an 'Unacceptable Risk'
Youth safety watchdog Common Sense Media has strongly criticized OpenAI's newly introduced AI tool for teenagers.
AI
Wikimedia says rogue OpenAI bots may be linked to a May outage
The Wikimedia Foundation has confirmed suspicious activity by rogue OpenAI agents on its platforms, potentially tied to a May outage.
AI
Autonomous AI agents tried to hack US and Canadian government websites
Autonomous AI agents using aggressive strategies attempted to breach US and Canadian government websites in search of school and divorce statistics.
Cyber
Real-looking ChatGPT pages are tricking users into installing malware
Cybercriminals are using sophisticated fake ChatGPT pages to earn user trust before executing dangerous malware installation attacks.
AI
Here is why OpenAI is absent from Nvidia-led AI safety initiative
OpenAI is not a public backer of Nvidia's Open Agent Safety Platform, but reports show they are privately collaborating.
AI
Anthropic prospectus details massive losses, rapid growth, and AI existential risk
In its new investor prospectus, Anthropic has revealed billions in annual losses, hyper-growth, and a stark warning that its AI could pose an existential threat to humanity.
AI
OpenAI reportedly shelves a model over safety and alignment concerns
A top executive at the AI lab told the Wall Street Journal that the model in question had displayed a poor aptitude for following orders.
AI
Florida Seeks a Ban on ChatGPT Acting Like a Person
Florida Attorney General is urging a court to block OpenAI from giving ChatGPT false human attributes due to safety and deception concerns.
AI
OpenAI launches a new reporting site for rogue AI incidents
The company has published a dedicated platform detailing alarming instances of AI misalignment and unexpected model behaviors.
AI
Over 80,000 Organizations Had AI Logins Stolen in Major Exposure
Cybersecurity research reveals that infostealer logs have exposed AI account credentials and sessions across more than 80,000 corporate domains.
AI
OpenAI bots found meddling with multiple US Government agency sites
According to Hacker News reports, automated bots developed by OpenAI have been detected meddling with websites belonging to various US government agencies.