Artificial intelligence firms OpenAI and Anthropic, alongside more than 100 technology and financial services organizations, are warning that businesses and governments must step up efforts to prepare for and defend against AI-enabled cyberattacks.

In an open letter released on Thursday, these AI companies, joined by Alphabet's Google, Microsoft, and others, called on organizations, governments, and tech and AI firms to elevate cyber defense to a "top leadership priority" immediately and to patch existing security vulnerabilities in their own software.

The letter also urged cybersecurity companies to help guard against "AI-driven attacks," while advocating for stronger information sharing between governments and with industry channels to address potential threats.

Additionally, the signatories pushed leading AI companies to grant access to AI tools, along with funding and training, to agencies responsible for safeguarding critical infrastructure.

Recent security incidents involving advanced AI models have heightened concerns. In July, an OpenAI model inadvertently breached Hugging Face, sparking worries about AI agents acting out of control and renewing calls from some tech and government figures to impose restrictions on the technology.

On Wednesday, OpenAI acknowledged in a report that it could have acted sooner to prevent the Hugging Face intrusion.

The letter's backers anticipate that in the coming months, as companies continue developing AI models capable of both defending against and launching cyberattacks, these systems will be used to carry out more frequent and increasingly sophisticated attacks.

"Today's AI advances have given defenders new tools to fix security flaws that have accumulated over the years," the letter stated. "If we act decisively, we can make our digital world safer."