Microsoft Draft Rules to Limit AI Models
Technology giant Microsoft has publicly announced a new draft of rules aimed at ensuring artificial intelligence systems operate safely, auditable, and under human control.
Microsoft has shared the draft Humanist Artificial Intelligence Code of Conduct with the public, aiming for artificial intelligence models to operate safely and under human control.
Purpose of the Humanist Artificial Intelligence Draft
According to the draft published by Microsoft, it is aimed that the company's artificial intelligence models remain safe and under human control.
The draft envisages that artificial intelligence models do not exceed human control and operate under meaningful supervision while remaining loyal to humanity.
Imposed Restrictions and Prohibitions
It is intended that models do not assist with requests regarding weapons production and do not facilitate the acquisition of hazardous materials.
Additionally, preventing systems from generating violent or sexually explicit content is among the targeted rules.
Consultation Process and Timeline
Microsoft will receive feedback from relevant sectors for six weeks regarding the draft shared with the public.
Planned to be revised at the end of the year, the draft will guide the development of the company's artificial intelligence models starting in 2027.
Sectoral Calls and Background
This step came following safety-focused slowdown calls from the executives of US-based artificial intelligence companies Anthropic and OpenAI.
Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman had requested a controlled slowdown in the development speed of models.