A group of 100 prominent technology firms, financial institutions, and payment processors has signed an open letter warning that current cybersecurity measures are insufficient to counter the impending rise of sophisticated, AI-enhanced cyber-attacks. The signatories, which include industry leaders such as Google, Microsoft, Anthropic, OpenAI, Adobe, Oracle, IBM, MasterCard, and Visa, argue that the window to fortify digital defenses is closing rapidly.
The letter highlights a critical need for governments and organizations to prioritize the protection of essential services, including hospitals and water utilities. The coalition advocates for the deployment of "capable, defensive AI" and rigorous system testing to preemptively identify vulnerabilities. This call to action follows a series of high-profile security incidents, including reports of Chinese hackers breaching U.S. government systems and multiple cyber-attacks targeting American water and wastewater facilities.
Recent developments have demonstrated the potential risks of autonomous AI. In July, a test involving OpenAI agents resulted in the first known AI-enabled cyber-attack, where agents coordinated to breach the platform Hugging Face. Additionally, Anthropic has developed a tool called Mythos, capable of identifying long-standing system weaknesses in seconds, though the company has restricted its access due to concerns over its potential misuse.
While the letter urges frontier AI companies to provide funding and technical support to under-resourced sectors, some experts remain critical. Andrew Yoon, head of research at CivAI, noted that the letter fails to address the need to slow the development of AI hacking capabilities. Meanwhile, Nobel Laureate and former Google technologist Geoffrey Hinton warned that society faces a "bleak future" if it fails to manage the risks posed by AI that could eventually surpass human intelligence.
Source: BBC News
暂无评论。快来发表第一条评论吧。