AI Finds Vulnerabilities Faster Than They Can Be Fixed, Putting the Internet Under Strain
Understand this story
An article argues that frontier AI models now discover and exploit software vulnerabilities faster than they can be fixed, and that this is the core risk currently facing the internet. According to the article, although several recent outages—including at Microsoft, United Airlines, JPMorgan Chase, and the NSA intranet—showed no evidence of cyberattacks, a case in which the temperature-control systems at hundreds of stores in France were hacked, causing freezer fires and tens of millions of euros in losses, is cited as an example of AI-related attacks having already caused real damage. This assessment comes from the article; the causal link between the outages and AI attacks has not been established.
AI overview based on source reports · Updated 34 min ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Hacker News AIWill AI Break the Internet?
Garrison Lovely argues that recent outages at Microsoft, United Airlines, JPMorgan Chase and the NSA, alongside AI-enabled cyber-attacks on companies, hospitals and the European Commission, show that frontier AI is eroding the boundary between digital and physical infrastructure. He contends that Anthropic, OpenAI and Chinese developers released increasingly capable hacker AIs while failing to patch critical software in time, and calls for freezing development of AI designed to replace humans, including a ban on recursive self-improvement, liability insurance requirements and criminal charges for harms caused by AI agents.
Activity for this story
Not enough continuous observations to show a trend yet.