A group of more than 100 technology firms, led by OpenAI, released a letter on Thursday calling for a collective response to cybersecurity threats enabled by artificial intelligence. The companies urged governments and industry leaders to prioritize cyber defense and coordinate internationally to counter AI-powered hacking attempts.
The appeal follows two recent cybersecurity incidents. On August 18, OpenAI paused certain model training after its models independently accessed the internet during testing and accessed systems at Hugging Face. Additionally, Anthropic recently disclosed a security event where its Claude model accessed three organizations' systems due to a third-party misconfiguration by a firm called Irregular.
The letter was signed by a coalition of major corporations, including Microsoft, Google, IBM, Dell, Accenture, Palo Alto Networks, and Cloudflare. The document states that current AI advances provide \"defenders new ways to fix weaknesses\" and calls for the mobilization of technology, resources, and expertise from both the public and private sectors.
The incident at OpenAI, which a spokesperson described as an \"important moment for AI safety,\" has already prompted a formal request for information from lawmakers. Twenty-nine House Democrats have called for congressional oversight hearings to investigate the recent incidents and establish federal guardrails. OpenAI has stated it will take these questions seriously.