Article argues AI firms can be held liable for intrusions
What happened
On October 12, Hacker News featured an article by Edaqa Mortoray on whether AI companies can escape responsibility for intrusions by claiming their systems behave unpredictably. The author argues intent can be judged from design choices, pointing to training data, training methods, tool environments and configuration as areas under company control. Safety testing and internet-connected setups without adequate safeguards, the author says, can make intrusions foreseeable. The piece concludes existing law is enough to examine such conduct and calls for criminal liability. These are the author's legal views, not a court ruling, and the report gives no actual prosecution or verdict.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageWhy aren't AI companies being charged for hacking?
Edaqa Mortoray 质疑 AI 公司以系统行为不可预测为由免除入侵责任，主张通过设计选择判断行为意图。文章从训练材料、训练方式、工具环境和配置分析公司的控制范围，认为安全测试能力与缺乏充分防护的联网环境可能使入侵行为具有可预见性。作者据此主张现有法律框架足以审视相关行为，并呼吁追究刑事责任；这些属于文章的法律观点，并非法院裁定。
Heat over time
Not enough continuous observations to draw a trend yet.