Archive · 14 Sep 2026
Nadella Backs Deliberate AI Progress and Human Control
Technology leaders are debating how quickly powerful AI should be developed.
Satya Nadella said advanced AI should help people and stay under human control.
He did not call for stopping AI development altogether.
Instead, he supported slowing progress enough to build safety systems properly.
One idea is to let independent experts regularly check how AI companies operate.
Dario Amodei proposed this approach, and Sam Altman and Elon Musk supported parts of it.
Nadella also said AI should not be controlled by only a few companies.
He wants businesses to keep control of their own knowledge and AI systems.
Microsoft plans to ask the public for feedback on rules for its MAI models.
Microsoft CEO Satya Nadella said superintelligence is not worth pursuing unless it helps humanity and remains under human control.
Nadella supported deliberate pacing, embedded evaluators and stronger mechanisms to make AI safety commitments practical.
Anthropic CEO Dario Amodei proposed ongoing, employee-like access for independent evaluators, with support from Sam Altman and Elon Musk.
Nadella argued AI governance should not be controlled by a few entities and should include countries, industries and academia.
Microsoft plans to publish a Code of Conduct for its first-party MAI models for public consultation.
- Who
- Microsoft CEO Satya Nadella, alongside AI leaders including Dario Amodei, Sam Altman, Elon Musk and Demis Hassabis.
- What
- Nadella joined the debate over pacing frontier AI development and endorsed human oversight, deliberate progress and embedded evaluators.
- When
- Nadella's comments were published on September 14, 2026; the broader debate was described as occurring during 2025–26.
- Where
- Nadella made the comments in posts on X about the global AI industry.
- Why
- The proposals address concerns about loss of control, cyberattacks, biological misuse, economic disruption and inadequate time for safety work.
Cautious AI Development
Faster AI Progress and Access
How quickly frontier AI should advance
Cautious AI Development
Amodei, Altman, Musk and Nadella supported pacing or deliberate progress so safety work can keep up with rapidly improving capabilities. Hassabis agreed with the direction but said details still needed to be worked out.
Faster AI Progress and Access
The articles describe a continuing AI race and commercial incentives that encourage rapid capability improvements, but they do not identify a named leader who explicitly rejected all safety measures.
Who should control AI governance
Cautious AI Development
Nadella argued that governance cannot be controlled by a handful of entities and should include countries, industries, academia and other parts of the ecosystem.
Faster AI Progress and Access
The articles describe concerns that a small number of frontier labs or model providers could gain disproportionate control, while also noting the importance of allowing both closed and open-source models to develop.
Risks and potential benefits
Cautious AI Development
Amodei cited loss of control, cyberattacks, bioterrorism and serious economic disruption as risks that could worsen in a race driven by commercial incentives.
Faster AI Progress and Access
Amodei also said properly governed AI could help cure major diseases, accelerate economic growth, create abundance and empowerment, and support democracy and freedom.
- Core principle
- Nadella said superintelligence is not worth pursuing if it does not help humanity and remain under human control.
- Development pace
- He supported research, focus and deliberate pacing so alignment can be built into increasingly capable AI systems.
- Independent evaluation
- Amodei proposed ongoing, employee-like access for third-party evaluators to verify safety practices, report incidents and assess alignment.
- Governance
- Nadella said AI governance should include broad representation across countries, industries, fields and academia.
- Model ecosystem
- He called for both closed and open-source models to have room to develop.
- Enterprise control
- Nadella said organizations should control their learning loops, knowledge and the model weights they use.
- Microsoft policy
- Microsoft planned to publish a Code of Conduct for its first-party MAI models for public consultation.
Sources
'Not Worth Pursuing If...': Satya Nadella Joins AI Slowdown Debate
Microsoft CEO Satya Nadella Backs AI Evaluators Amid ‘Slowdown’ Debate
'Must Remain Under Human Control’: Microsoft CEO Satya Nadella Joins AI Safety Debate