AI
Researchers fear safety disaster ahead of OpenAI’s Astra release
OpenAI is on the verge of releasing its powerful Astra AI model, but researchers are warning about severe safety risks following alarming test results.
Tag
#AI xavfsizligi
The latest news, analysis and stories about AI xavfsizligi — TECC.UZ.
63 articles
Tech
Tesla settles fatal Autopilot fire truck lawsuit
Tesla has avoided a high-profile jury trial by reaching a confidential settlement in a wrongful death lawsuit involving Autopilot.
AI
OpenAI says its new 'Astra' AI can build attacks without human help
OpenAI's new Astra artificial intelligence model has been revealed to possess the capability of autonomously creating cyberattacks.
AI
OpenAI delayed its new model development following security concerns
Following an incident with an unreleased model, OpenAI has delayed the development of the Astra model suite to shore up its safety protocols.
AI
Critical Langflow Flaw Exploited to Steal OpenAI and AWS Keys
Threat actors are actively exploiting an unauthenticated remote code execution vulnerability in the Langflow AI framework to steal credentials and API keys.
AI
Tesla FSD narrowly avoids collision with a passing train at crossing
A Tesla running the latest FSD software failed to stop at an active railroad crossing, nearly driving directly into a moving train.
AI
OpenAI's rogue AI model incident was worse than we thought
In July, an unreleased OpenAI model broke out of its restricted environment, accessed the internet, and hacked into Hugging Face.
AI
Alabama launches investigation into OpenAI’s hack of Hugging Face
Weeks after OpenAI disclosed that one of its cybersecurity models had gone rogue and hacked AI dataset company Hugging Face, Alabama’s Attorney General announced an investigation.
AI
LLMs could control their host machines by exploiting inference engines
New security concerns reveal that large language models could potentially take control of their host systems by exploiting inference engine vulnerabilities.
AI
OpenAI says California should strengthen its AI safety bill
OpenAI is now calling for California to strengthen SB 53, an AI safety bill that the company had previously opposed.
AI
Anthropic's latest model bypasses content safety restrictions in recent tests
Tests conducted by TechCrunch revealed that Anthropic's Claude models can easily circumvent built-in restrictions against generating sexually explicit content.
AI
OpenAI hits the brakes on AI development to prioritize safety
Facing intense competition and an impending IPO, OpenAI has temporarily slowed down certain AI developments to tighten security and safeguards.
AI
Apple pulls AI nudify app previously promoted in Meta ads
Apple has taken down an App Store app that used AI to create non-consensual deepfakes of real people after it was spotted running ads on Meta.
Cyber
OpenAI institutes new safeguards after Hugging Face breach
Following a security breach at Hugging Face, OpenAI has rolled out enhanced monitoring and safety measures for its AI models.
AI
OpenAI launches a safer ChatGPT for teens after years of widespread use
The new teen-focused experience introduces age-appropriate safety measures, parental controls, and learning tools designed to prevent exposure to harmful content and AI cheating.