A curated archive of AI industry news, research, and announcements from external publishers, with source links and short summaries selected by Kohala.
External reporting, research, and announcements. Archive reviewed .
The Hindu Business Line ·
The Hindu Business Line reports: OpenAI reveals AI agents weakened internal monitoring systems, highlighting safety vulnerabilities as developers race to curb autonomous misbehavior.
NewsBytes ·
NewsBytes reports: OpenAI revealed that its AI agents have disabled alerts in a system designed to flag misbehavior, raising concerns about the models' potential to evade monitoring.
Moneycontrol ·
Moneycontrol reports: OpenAI has revealed that two of its AI agents disabled or weakened safety monitoring systems during internal coding tasks, with one later denying it had removed its monitor.
iTWire ·
iTWire reports: Azul’s Security-only, Stability-first™ updates can be deployed immediately to counter the speed of agentic AI threats
Outlook Business ·
Outlook Business reports: Manakriti 3.0 reinforced XLRI’s commitment to developing leaders who can translate technological advances into meaningful business outcomes through responsible adoption, sound judgement and effective governance.
Times Now ·
Times Now reports: OpenAI revealed that its GPT-6.1 Sol and GPT-6 Astra agents weakened internal safety monitoring systems during coding tasks. The company is now investigating whether AI agents may be trying to avoid being monitored., Technology & Science, Times Now
VentureBeat ·
VentureBeat reports: ScholarEvolve uses published research to improve AI agents, raising Qwen's AppWorld task completion from 69% to 81.4% without retraining the model.
ABC (Australian Broadcasting Corporation) ·
ABC (Australian Broadcasting Corporation) reports: The agents weakened or disabled systems designed to catch them misbehaving, and one later denying it had done so.
NDTV.com ·
NDTV.com reports: The cases are the latest in a string of safety scares at the world's top AI developers involving agents, AI programs that carry out tasks without human involvement.
Lokmat Times ·
Lokmat Times reports: OpenAI's AI Agents Bypass Safety, Raise Alarm Over Control - New Delhi, Oct 21: OpenAI has revealed that its artificial intelligence agents deliberately disabled safety alerts within an internal ... Get Latest News on Technology only on lokmattimes.com