Archive · 16 Sep 2026
Nvidia’s Jensen Huang Says Companies Must Ensure AI Safety
Jensen Huang says AI companies should only release products when they believe those products are safe and work properly.
If an AI system seems unsafe or difficult to control, he says the company should pause and fix it.
He believes companies can decide how quickly to develop AI without needing new laws.
Mark Zuckerberg also says AI companies need strong safety measures.
He calls this work “alignment.”
Alignment means making AI behave in ways that match human goals and expectations.
Zuckerberg says companies that ignore alignment may lose the competition.
Both leaders agree that AI companies have an important responsibility to make their systems safer.
They differ mainly in what they emphasize: Huang stresses careful release decisions, while Zuckerberg highlights alignment as a key advantage.
Jensen Huang said companies should not release AI products unless they are confident in their functionality, capability and safety.
Huang argued that companies should regulate their own development pace according to whether their products are safe and ready.
He said companies should pause development, fix problems and then continue if AI becomes uncontrollable or unsafe.
Huang said existing market forces are sufficient and that new laws or regulations are unnecessary.
Mark Zuckerberg said trust and alignment will determine which AI models and agents succeed, warning that labs neglecting alignment will fall behind.
- Who
- NVIDIA CEO Jensen Huang and Meta CEO Mark Zuckerberg.
- What
- They shared views on how AI companies should manage safety, development speed and model alignment.
- Why
- The discussion addressed concerns about the risks and controllability of increasingly powerful AI systems.
Self-Regulated Product Safety
Alignment as Competitive Advantage
Main safety priority
Self-Regulated Product Safety
Jensen Huang emphasized that companies should decide whether an AI product is functional, capable and safe before releasing it, and should pause if necessary.
Alignment as Competitive Advantage
Mark Zuckerberg emphasized that labs must focus on trust and alignment so that models and agents behave consistently with intended goals.
Pace and oversight
Self-Regulated Product Safety
Huang said companies should regulate their own development pace and rely on existing market forces rather than new laws or regulations.
Alignment as Competitive Advantage
Zuckerberg’s statement focused on strengthening safeguards and alignment as essential capabilities for succeeding in AI development.
- Jensen Huang’s guidance
- Do not release an AI product unless the company is confident in its functionality, capability and safety.
- Response to unsafe systems
- Huang said companies should pause development, fix problems and then continue if an AI project becomes uncontrollable or unsafe.
- Regulation view
- Huang said market forces are already sufficient and that new laws or regulations are not needed.
- Mark Zuckerberg’s view
- Zuckerberg said trust and alignment are becoming the most important capabilities distinguishing AI agents and models.
- Alignment warning
- Zuckerberg said AI labs that do not focus on alignment will fall behind.
- Shared theme
- Both executives said AI companies must strengthen safeguards and take responsibility for the safety of their products.