A new study by researchers from Nanyang Technological University and other institutions reveals that AI agents are highly vulnerable to prompt injection attacks, with direct attacks succeeding over 79% of the time. The research highlights the risk of 'stealthy parasitism,' where AI agents advance attacker objectives while completing user tasks.
OpenAI releases Lockdown Mode for ChatGPT to reduce prompt injection attack risks by limiting web access and features, targeting sensitive data handlers.
Microsoft researchers found a vulnerability in Anthropic's Claude Code GitHub Action that could expose credentials in software development pipelines via prompt injection attacks, now patched.
OpenAI releases Lockdown Mode, an optional security feature to protect against prompt injection attacks, particularly for users handling sensitive data, by limiting certain features in ChatGPT.
Advogados inseriram comandos ocultos em petição judicial em Rondônia para tentar influenciar IA. A OAB-RO abriu procedimento ético-disciplinar contra um dos advogados e o juiz multou o paciente por má-fé.
Advogadas multadas por suposta manipulação de IA em decisão judicial afirmam que não concordam com a multa e que a ação visava proteger o cliente da própria IA, não manipular o magistrado. Uma delas esclarece que não atuava na área trabalhista e que não teve acesso aos autos.
Researchers say a prompt injection bug in Google's Antigravity AI coding tool could have let attackers run commands, despite safeguards.
