STORY · FORSKNING_
OpenAI shows that reasoning models have poor control over chains of thought – and thinks that's a good thing
OpenAI presents a new tool called CoT-Control and documents that advanced reasoning models struggle to control their internal chains of thought. The researchers conclude that this lack of control actually makes the models easier to monitor and understand.
WHY IT MATTERS
The ability to control and inspect reasoning processes is critical for AI safety. If even the models themselves cannot fully control their chains of thought, it becomes easier to detect manipulative or hidden behavior.
SOURCES
MACHINE-GENERATED SUMMARY This summary is written by machine from the sources below. We sort and explain — but we are a way into the field, not the final word. Check the source when something matters to you.