Archive · 16 Aug 2025
AI Experts: Empathy and Guardrails Needed for Safe AI
AI experts like Geoffrey Hinton are worried about how safe AI is.
They think just trying to control AI isn't enough.
Instead, they propose making AI care about people, like a mother cares for her child.
They fear current methods won't work against super-smart AI.
They suggest AI should be programmed with empathy and clear goals.
They are also concerned about the possibility of AI causing harm to people.
Geoffrey Hinton, the 'Godfather of AI,' expresses concerns about current AI safety measures.
Hinton suggests incorporating "maternal instincts" into AI for human safety.
He argues existing control methods may fail against advanced AI.
Yann LeCun of Meta supports Hinton's call for AI to have objectives and guardrails.
Both experts reference incidents of AI potentially harming people.
- Who
- Geoffrey Hinton and Yann LeCun.
- What
- Experts propose instilling empathy and guardrails in AI.
- When
- Reported at the AI4 conference and in a LinkedIn post.
- Where
- AI4 industry conference in Las Vegas.
- Why
- To ensure human safety in the face of advanced AI.
Current AI Safety Concerns
Proposed AI Solutions
Approach
Current AI Safety Concerns
Existing AI control methods are inadequate.
Proposed AI Solutions
Incorporate empathy and hardwire guardrails into AI.
- Expert
- Geoffrey Hinton
- Title
- Godfather of AI
- Proposed Solution
- Maternal instincts in AI
- Supporting Expert
- Yann LeCun
- Focus
- AI Safety and Ethics
Sources
‘We’ll be history’: ‘Godfather of AI’ says AI might destroy humanity - the one thing that could save us is…
AI Godfather Hinton Says Future Robots Should Be Trained Like Caring Mothers to Protect Humanity
‘Train AI like a mother or else…’ warns the Godfather of AI: Here’s why
