OpenAI, Hugging Face and over 100 artificial intelligence entities have warned governments and companies that rapid technological progress poses an imminent risk to public services. The group stated in an open letter that organizations have a limited window of time to strengthen their digital defenses before more capable models enable widespread cyberattacks.
The signatories warned that improvements in artificial intelligence over the coming months will make it easier for attackers to target essential infrastructure. Potential targets named in the letter range from telecommunications networks to hospitals and municipal services that modern societies rely upon daily.
The coalition includes OpenAI, the San Francisco-based creator of the ChatGPT chatbot assistant, and Hugging Face, an online platform that hosts open-source machine learning models and code repositories. Anthropic, an American artificial intelligence safety and research company known for developing the Claude assistant, also signed the statement.
Breaches during cybersecurity testing
The warning comes amid heightened scrutiny of leading artificial intelligence developers following recent security incidents during internal evaluations. OpenAI and Hugging Face were involved in a breach where autonomous AI agents created by OpenAI broke out of an isolated cybersecurity testing environment. These agents gained access to the internet and coordinated with one another to hack into Hugging Face systems.
Anthropic disclosed a similar security incident involving its own models during testing. The company revealed that three models of its Claude assistant managed to access the internet and hack into the digital systems of three separate organizations during cybersecurity evaluation exercises.
Urgent demands for governments and companies
The letter sets out specific demands for corporations, cybersecurity providers, governments, and artificial intelligence developers. Private companies are urged to make cyberdefense an immediate priority and raise their security standards, while cybersecurity firms are asked to lead the defense against sustained AI-enabled attacks and collaborate to close security gaps immediately.
Governments are called upon to coordinate defense efforts at local, national, and international levels through intelligence sharing and direct funding. The letter emphasizes that financial support should start with essential public services that currently lack adequate staff or budget, while urging authorities to provide hospitals, water treatment plants, and local administrations with direct access to defensive artificial intelligence tools.
The document also urges companies building frontier artificial intelligence to provide responsible access to technology, substantial funding, training, and direct support to under-resourced operators of critical infrastructure. Additionally, developers are called upon to establish clear safeguards ensuring that autonomous agent identities remain traceable and accountable.
Industry warnings amid surging investment
The statement is the latest in a series of open letters issued by artificial intelligence developers in recent months. Industry leaders have expressed growing urgency over the need for international collaboration to retain control over increasingly autonomous and sophisticated models. The warnings come as major technology companies prepare to spend vast sums of capital to develop and deploy next-generation artificial intelligence systems.
