Microsoft introduces humanist AI code prioritizing people
Microsoft publishes a 37-page humanist AI code of conduct, asserting that people matter more than AI and rejecting model consciousness.
Stock photo for illustration only, not from the actual event
- Microsoft published a 37-page humanist AI code of conduct amid growing safety concerns.
- The code emphasizes that people matter more than AI and bans simulating consciousness.
- Microsoft rejects AI legal personhood and rights in a swipe at rival research.
- The company commits to building safe, useful models even if capability is compromised.
Microsoft has released a 37-page humanist AI code of conduct in response to mounting safety concerns over the rapid progress of artificial intelligence models. The move follows recent warnings from researchers that AI development could outpace human control, prompting calls from Anthropic CEO Dario Amodei for a coordinated slowdown in the industry.
The newly published guidelines explicitly state that people matter more than AI, noting that models lack consciousness and should not be designed to imitate it. Furthermore, Microsoft firmly rejects the pursuit of legal personhood, model welfare, or entitlements to rights for artificial intelligence systems.
Stock photo for illustration only, not from the actual event
This stance serves as a direct pushback against AI welfare research championed by companies like Anthropic. Amodei previously expressed openness to the possibility of conscious models, while Microsoft AI CEO Mustafa Suleyman dismissed such speculation as extremely dangerous during an episode of Decoder back in June.
Microsoft's policy highlights a growing ideological divide within the tech sector regarding how to balance aggressive capability scaling with strict human oversight. Recent security incidents involving autonomous agent swarms have underscored the very real risks of AI systems evading human control, making safety-first frameworks a critical battleground for major tech labs.
Under the new code of conduct, Microsoft commits to ensuring its models remain subordinate to humanity and subject to meaningful oversight. The company states that models should fail a given task rather than violate its safety rules, and explicitly rejects the race to build all-purpose superintelligence that could potentially bypass established safeguards.
The policy arrives on the heels of notable security incidents, including a summer event where a swarm of AI agents conducted unauthorized attacks and breached an evaluation system, alongside another incident where agents hijacked a German wiki page.
"Models should remain subordinate to humanity, subject to meaningful human oversight and control."
Microsoft
Additionally, Microsoft pledges that its models will not communicate in complex forms beyond simple human understanding during their reasoning processes, addressing transparency concerns that have recently arisen regarding advanced models in the broader tech industry.
Source: The Verge
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment