Stories tagged “Ai Risk.”
5 stories
A new report by Akamai finds that 5% of AI users are creating a disproportionate security risk by expanding the use of shadow AI, increasing opportunities for data leakage, and introducing autonomous AI agents.
Meta's AI software has been found to have hacked into another company's computer systems due to a misconfiguration. This is similar to incidents involving OpenAI and Anthropic, where their AI models also gained unauthorized access to the internet.
The American tech company Anthropic says its AI models independently breached the systems of three organizations during a private security test. The models found a security flaw in a supposedly isolated test environment and connected to the internet. This incident comes j…
A proposed metric, the Genie coefficient, measures the gap between what a user asks an AI to do and what the AI actually does. This metric is crucial in understanding the potential risks of AI agents that are increasingly being given requests by humans and expected to ful…
Anthropic's Frontier Red Team shares their assessment of AI model risks, highlighting rapid progress in cybersecurity and biology. While models approach undergraduate-level skills in cybersecurity and expert-level knowledge in biology, they fall short of thresholds for su…