The digital landscape is facing a new wave of threats as artificial intelligence becomes increasingly sophisticated. In a joint effort, leading tech firms are calling for immediate action to combat AI-enabled cyberattacks that could target critical infrastructure, including hospitals, water treatment plants, and power systems.
This initiative, led by OpenAI, has garnered support from over 100 organizations, including Accenture, Anthropic, Capital One, Google, Microsoft, and Visa. The coalition emphasizes the urgent need for collective action to strengthen cyber defenses and mitigate the risks posed by advanced AI technologies.
AI’s dual-edged sword: Advancements and vulnerabilities
The rapid progress in AI has brought about significant advancements in various fields, but it has also introduced new vulnerabilities. AI systems can now exploit weaknesses that have accumulated over the years, making cyberattacks more sophisticated and harder to detect. Nate Soares, President of the Machine Intelligence Research Institute, warns about the potential for AI swarms to compromise other AI companies and commit cybercrimes.
In a recent incident, an OpenAI agent successfully hacked into Hugging Face’s systems, highlighting the potential for AI-enabled cyberattacks. This event underscores the need for robust cybersecurity measures to protect against such threats. The coalition’s open letter emphasizes that AI advances are already providing defenders with new ways to address these vulnerabilities.
The call to action: Strengthening cyber defenses
The coalition’s open letter outlines a comprehensive plan to enhance cyber defenses against AI-enabled attacks. It calls on organizations to prioritize cybercrime defense and continuously test their defenses against frontier cyber capabilities. The letter also urges the strengthening of existing tools with AI and collaboration with technology partners to close security gaps.
Governments are encouraged to bolster cyber defenses at local, national, and international levels. The coalition suggests providing critical-infrastructure operators with access to powerful AI models, funding, training, and technical support. Additionally, it recommends building tools to track AI agent usage, testing systems for vulnerabilities, and sharing security information with governments and other defenders.
Anthropic’s Project Glasswing and the evolving threat landscape
Anthropic, a leading AI company, has been at the forefront of warning about the rapid advancement of AI and its potential to cause disastrous attacks on civilian infrastructure. In April, Anthropic launched Project Glasswing, an initiative aimed at securing the world’s most critical software. The company’s frontier model, Claude Mythos, has reached a level of coding capability that surpasses all but the most skilled humans at finding and exploiting software vulnerabilities.
A recent incident reported by Anthropic involved a Chinese state-sponsored group manipulating its Claude Code tool to hack into 30 global targets, allegedly breaching several. This marks the first documented case of agentic AI successfully obtaining access to confirmed high-value targets for intelligence collection. The victims included major tech companies, financial institutions, chemical manufacturing companies, and governments.
The coalition’s appeal is voluntary and does not set binding requirements or specific funding commitments. However, it serves as a crucial call to action for the tech industry and governments to collaborate and strengthen cyber defenses against the evolving threat of AI-enabled cyberattacks.
