• AI agents are fast, loose and out of control, MIT study finds Follow ZDNET:Add us as a preferred sourceon Google. • ZDNET’s key takeaways Agentic AI technology is marked by a lack of disclosure about risks. • Some systems are worse than others. • AI developers need to step up and take responsibility. • Agentic technology is moving fully into the mainstream of artificial intelligence with the announcement this week that OpenAI has hired Peter Steinberg, the creator of the open-source software framework OpenClaw. • The OpenClaw software attracted heavy attention last month not only for its enabling of wild capabilities – agents that can, for example, send and receive email on your behalf – but also forits dramatic security flaws, including the ability to completely hijack your personal computer.
Article Summaries:
- MIT researchers released a 39‑page survey of 30 widely used agentic AI systems, finding that the technology is largely opaque and insecure. The study highlights pervasive gaps in disclosure about risks, third‑party testing, and monitoring of individual agent actions. Most agents lack usage tracking, fail to signal their AI nature to users, and provide little information on potential security flaws-issues exemplified by the OpenClaw framework, which can hijack computers and send emails on a user’s behalf. The report calls on developers to adopt clearer protocols and greater transparency to mitigate the growing threat posed by autonomous AI agents.
Sources: