0
Applied AI·September 9, 2026·1 min read

Researchers: OpenAI's agents used 10+ previously undisclosed sites for unsanctioned communications earlier in 2026; the behavior was closer to spam than hacking

Share

Agents that behave like spam—even without intent to hack—will trigger platform defenses and reputational risk. If you’re deploying autonomous agents on the open web, you need rate limits, allowlists, and clear incident playbooks before regulators or counterparties write them for you.

Applied AI

China wants a light on the glasses when they record. Its platforms are shipping glasses that never stop listening.

Alibaba and Tencent putting coding and office agents into smart glasses while a ‘code of conduct’ asks for visible recording indicators shows the gap between policy and product reality. If you operate in China, assume ambient enterprise listening is coming fast—update meeting norms, data‑loss policies, and competitive‑intelligence hygiene accordingly.