Microsoft · Safety and security · Agents
Microsoft CEO Satya Nadella proposes seven principles for treating frontier AI models as insider risks
Microsoft CEO Satya Nadella published an essay titled “Models as Insider Risks in the Super Intelligence Era” on the personal blog sn scratchpad on October 10 and shared it on X. Nadella argued that companies should treat closed and open-weight frontier models like insider risks, “not because they are necessarily malicious,” but because any capable actor with access to important systems can make mistakes or be compromised. According to the essay, a model provider’s assurances do not remove responsibility, and the controls over what a model can access and do “must sit outside the model.”
The seven principles are model diversity (no model as the sole dependency or its own verifier), observe everything (tamper-proof, human-readable evidence of every meaningful action), verifiability (continuous testing, including failures and attacks), independent controls, independent auditability, containment (an authorized person can pause or shut down a model mid-task) and incident disclosure (timely notice and shared lessons). The essay announces no Microsoft product or policy.
Sources
- sn scratchpad (Satya Nadella), “Models as Insider Risks in the Super Intelligence Era”, Original essay on Nadella's personal blog with the full text of the seven principles, (snscratchpad.com)
- Business Today, “Satya Nadella flags ‘black box’ AI risks: Why Microsoft CEO wants stronger controls over superintelligence”, Report on the essay as shared on X, (businesstoday.in)
- The Next Web, “Satya Nadella says companies should treat AI models as insider risks”, Summary of the principles and reactions on X, (thenextweb.com)
About this story
Spotted an error in this story? [email protected] · Instagram
Short link: thejarrus.com/en/nadella-insider-risk
This story in Turkish: Microsoft CEO’su Satya Nadella, öncü yapay zeka modellerinin şirketlerde içeriden risk gibi ele alınması için 7 ilke önerdi
On the same topic
Anthropic says small AI models may be months away from being able to self-replicate on IoT devices
An Anthropic Frontier Red Team report says small models may soon be able to copy themselves onto devices such as smart ovens and Android TV boxes.
Anthropic says its new distillation defenses have cut the data harvested from Claude by around 90 to 99 percent
Anthropic says preserved thinking, restricted reasoning access and a two-stage classifier have cut the data distillers harvest from Claude by 90-99%.
Linus Torvalds built an experimental Linux kernel patch series with Claude Code
Torvalds says Claude Code wrote the code, tests and first draft. One test ran 23 times faster, but the series isn’t merged and was handed to Kiryl Shutsemau.