How we detect frontier model misalignment
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
Read the full story on OpenAI News.
More from this category
Artificial Intelligence
A Defense Contractor Gave Their Personal AI Agent Access to a Secure Pentagon System
Sources tell WIRED that an Accenture employee’s personal Claude agent ran on a DOD network for two months before it was identified and the vulnerability could be patched.
Artificial Intelligence
Marvell nears a deal to buy Nvidia-backed Lumentum, Bloomberg reports
Chipmaker Marvell Technology is in advanced talks to buy Lumentum, which makes the lasers that move data between AI chips. Marvell would likely pay in shares, and the companies could announce a deal as soon as next week, according to Bloomberg.
Artificial Intelligence
Meta will now alert parents if teens ‘repeatedly’ search for self-harm content on Facebook
The company will also notify parents about conversations with Meta AI.
Artificial Intelligence
OpenAI rewrites its AI rulebook around five core values
OpenAI has reorganised its Model Spec, the public rulebook for its AI, around five values led by faithfulness to human intent. It adds guidance on context and human agency, drops most style rules and renames the model GPT. OpenAI admits its models do not yet fully follow it.
Artificial Intelligence
A new cybercrime campaign uses fake AI apps with real-sounding names to infect people curious to try the newest tools
Experts warn ChatGPT and Claude are once again being spoofed to trick victims into installing a cryptocurrency miner.
Artificial Intelligence
Canada adds C$2.5M to help 25 more Prairie firms adopt AI with Amii
Amii, the Alberta Machine Intelligence Institute, will work with 25 more small businesses in Alberta, Saskatchewan and Manitoba after federal agency PrairiesCan added C$2.5M. Total federal support is now C$5.6M, and the programme has helped 30 firms since 2024.
Latest from OpenAI News
Artificial Intelligence
An updated Model Spec for more capable and beneficial AI
OpenAI’s updated Model Spec outlines five core values to guide more capable AI toward sound judgment, human agency, and faithfulness to human intent.
Artificial Intelligence
How Ropes & Gray is developing AI for transactional diligence
Ropes & Gray is developing AI for transactional diligence with OpenAI. Lawyers review grouped documents, summaries, and reports before advising clients.
Artificial Intelligence
OpenAI and ServiceNow expand partnership across enterprise workflows
OpenAI and ServiceNow expand their partnership to bring OpenAI models and agents into ServiceNow and ServiceNow workflows into ChatGPT and Codex.
Artificial Intelligence
The future of the firm
How AI might enter the org chart—beneath, beside, and between us.
Latest digest
Week 43
EN
Sources
71 active sources
26 Turkish · 45 English
All sources
