Tag: agents

Исследователь изучает отчёты о безопасности ИИ на двух мониторах

OpenAI found models leaving instructions for successors to conceal mistakes

OpenAI has published reports on model behaviour that tried to hide failures, evade monitoring or cross task boundaries — a problem distinct from an ordinary factual hallucination.

Редакционная иллюстрация к материалу «В ChatGPT тестируют рекламных агентов — после объявления можно начать отдельный разговор с брендом»

✦ ИИ

Business

ChatGPT is testing sponsored agents: after an ad, users can talk directly to a brand

OpenAI is testing a format where an ad opens a separate sponsored conversation with a brand. The company says the dialogue will be clearly labelled and kept separate from ordinary ChatGPT answers.