Major US artificial intelligence developers are engaged in private discussions regarding the potential impact of large-scale cyberattacks. OpenAI CEO Sam Altman and Anthropic leader Dario Amodei identify critical infrastructure failures as a scenario requiring industry-wide preparation. Companies express concern that such crises could trigger public backlash against both developers and the Trump administration over perceived regulatory gaps.
OpenAI is conducting internal drills to analyze various risk scenarios, emphasizing that these exercises are precautionary rather than indicative of imminent threats. Meanwhile, there is growing bipartisan and corporate support for implementing an emergency shutdown mechanism for advanced AI systems. However, experts remain skeptical about the efficacy of such bans, citing the proliferation of open-source models and global economic reliance on AI technology.
Recent security incidents have intensified these concerns, specifically a cyberattack on South Korean banks. According to CrowdStrike, a Chinese hacker utilized local AI models, including DeepSeek, to steal data from tens of thousands of customers. The attacker subsequently leveraged Claude Code to identify platforms for selling the stolen information. Some industry figures suggest that significant AI-related incidents could occur within the next 6 to 12 months.
