Archive · 12 Jun 2025
AI Pioneer Creates New System To Make AI Safer, Prevent Harm
Some people are worried that AI could be used for bad things, like making bombs.
One of the top AI experts, Yoshua Bengio, is creating a new AI called "Scientist AI" to make AI safer.
This new AI will be able to explain its answers and say how sure it is.
It will also have a "world model" to better understand how things work, like how hands move.
This could help it avoid mistakes.
Bengio hopes this new AI can help keep other, potentially dangerous AI systems in check.
FBI revealed AI was used to obtain bomb-making instructions.
Yoshua Bengio launched LawZero, a nonprofit for safer AI.
Bengio's "Scientist AI" will assess confidence and explain reasoning.
Scientist AI may act as a guardrail against unsafe AI systems.
Scientist AI incorporates a "world model" for improved understanding.
- Who
- Yoshua Bengio, along with his nonprofit LawZero, is developing the new AI model.
- What
- New AI model called "Scientist AI" is being developed to improve AI safety.
- When
- Launch of LawZero and the development of Scientist AI happened around the time the FBI revealed AI was used for bomb-making instructions.
- Where
- N/A
- Why
- To make AI safer and prevent it from being used for harmful purposes.
Current AI systems
Scientist AI
Focus
Current AI systems
Prioritize speed and new approaches
Scientist AI
Prioritize safety and explainability
- Organization
- LawZero
- Model Name
- Scientist AI
- Key features
- Assesses confidence, explains reasoning, incorporates a world model
- Primary Goal
- To create a safer AI system.