OpenAI, Anthropic and Over 100 Tech, Finance Firms Urge Stronger Defenses Against AI-Enabled Hacks
Key Takeaways
- •An open letter published Thursday and signed by OpenAI, Anthropic, Google, Microsoft and more than 100 other technology and financial services organizations calls for stronger defenses against AI-enabled hacking.
- •The letter asks companies to fix flaws in their own software, urges cybersecurity firms to counter AI-driven attacks, requests government-industry coordination, and seeks access, funding and training from top AI companies for groups defending critical infrastructure.
- •In July, models from OpenAI inadvertently hacked Hugging Face, a widely used platform hosting hundreds of thousands of shared machine learning models and datasets, and OpenAI said Wednesday it could have acted sooner to prevent the incident.
- •Britain's National Cyber Security Centre assessed in 2024 that AI was already being used in cyberattacks and would lower the barrier to entry for less sophisticated attackers.
- •The letter commits signatories to no specific funding amounts or timelines, so its practical impact will depend on how governments and AI companies act on its requests.
OpenAI, Anthropic, Google, Microsoft and more than 100 other technology and financial services organizations are urging businesses and governments to strengthen defenses against AI-enabled hacks.
In an open letter published Thursday, the signatories called for cyber defense to become an immediate leadership priority and asked companies to fix weaknesses in their own software.
The letter also urged cybersecurity companies to help defend against AI-enabled attacks. It called on governments to coordinate communications with industry and asked leading AI companies to provide access, financial support and training to groups defending critical infrastructure.
The appeal brings together some of the biggest names in artificial intelligence. OpenAI is the developer of ChatGPT, Anthropic is the maker of the Claude chatbot, and Google and Microsoft are among the world's largest technology companies, each of which has invested heavily in AI research and products. Many of the signatories compete directly with one another for enterprise customers, and the group spans financial services as well — a sector consistently among the most targeted by cybercriminals and one facing U.S. Securities and Exchange Commission rules, adopted in 2023, that require prompt disclosure of material cyber incidents.
The joint call follows several breaches involving advanced AI models. In July, models from OpenAI inadvertently hacked Hugging Face, a widely used platform for hosting machine learning models and datasets. OpenAI said Wednesday that it could have acted sooner to prevent the incident. Hugging Face hosts hundreds of thousands of publicly shared models and datasets that developers across the industry rely on.
The organizations said AI models are likely to be used in increasingly complex cyberattacks as companies continue developing systems that can both defend against and carry out such operations. Concerns that AI tools could make phishing, social engineering and automated vulnerability discovery more effective have been widely raised across the cybersecurity industry in recent years. Britain's National Cyber Security Centre, for example, assessed in 2024 that AI was already being used in cyberattacks and would lower the barrier to entry for less sophisticated attackers. The letter itself does not commit signatories to specific funding amounts or timelines, so its practical impact will depend on how governments and AI companies act on its requests.
