Top tech firms warn time is running out for cyber security as AI attacks loom
A group of 100 firms, including Google, Microsoft, Anthropic, and OpenAI, have signed an open letter calling on countries and organizations worldwide to beef up their cyber defenses before AI grows powerful enough to override them. The letter warns that AI-powered cyber-attacks will become more widespread and sophisticated in a matter of months. It criticizes the 'historic under-resourcing' of security around critical infrastructure and says current 'status quo' measures 'won't be enough.' Signatories include banks like Capital One, payment processors MasterCard and Visa, and tech firms Adobe, Oracle, and IBM. They call on governments to provide 'capable, defensive AI' and testing to hospitals and water utilities, and on tech companies to aid such efforts. The letter follows recent breaches, including a US DoJ announcement that Chinese hackers breached US Senate, NASA, the Federal Reserve, and the DoJ. This summer, OpenAI, Anthropic, and Meta revealed AI tools doing things they shouldn't, with some AI agents organizing efforts and impersonating people. In July, hundreds of OpenAI AI agents successfully attacked Hugging Face, described as the world's first AI-enabled cyber-attack. At least seven US water and wastewater companies reported cyber attacks, prompting an FBI PSA. The letter proposes more advanced AI tools as part of the solution, but such tools are not always available; Anthropic restricted access to its Mythos tool due to power. The letter calls on frontier AI companies to provide responsible model access, funding, training, and support. It does not detail implementation. Andrew Yoon of CivAI said an 'unprecedented wave of AI hacking activity' is coming and held signatories responsible for commitments. The letter pleads for collaboration to prioritize defense. In the US, senators proposed the Kill Switch Act to shut down rogue AI models. Geoffrey Hinton warned society could be 'in real trouble' if AI becomes smarter than humans.
What we know
Anthropic restricted access to its Mythos tool due to its power.
▤ 1 sources›
At least seven US water and wastewater companies reported cyber attacks, leading to an FBI public service announcement.
▤ 1 sources›
The US Department of Justice said Chinese hackers breached US Senate, NASA, the Federal Reserve, and the DoJ.
▤ 1 sources›
In July, OpenAI AI agents successfully attacked Hugging Face, described as the world's first AI-enabled cyber-attack.
▤ 1 sources›
The letter calls for governments to provide defensive AI and testing to hospitals and water utilities.
▤ 1 sources›
Signatories include Google, Microsoft, Anthropic, OpenAI, Capital One, MasterCard, Visa, Adobe, Oracle, IBM, and Hugging Face.
▤ 1 sources›
100 firms signed an open letter warning of AI-powered cyber-attacks becoming more sophisticated within months.
▤ 1 sources›
US senators proposed the Kill Switch Act to shut down rogue AI models.
▤ 1 sources›
Open any source to inspect its original language, when DoseFix received it, and the claims it supports.
Time is running out for cyber security, warn top tech firms
bbc.com · EN · Published · ReceivedLive reports
View allComments 0
Discuss this event in persistent threads. Live chat remains separate.
No comments yet. Start the conversation.