Microsoft has published a new code of conduct for its artificial intelligence models, placing human control and safety at the centre of how its systems should operate.
The code, released by Microsoft’s AI division, is based on what the company describes as “Humanist AI” — an approach designed to ensure that AI systems remain under human control rather than operating independently.
Microsoft said the central principle behind the framework is that people should remain more important than AI.
The move comes amid growing concerns over the potential risks posed by increasingly capable AI systems, including their use in cyberattacks.
Microsoft cited recent coordinated hacking campaigns involving AI agents as evidence of the need for stronger safeguards across the industry.
The company’s initiative also follows calls from some leading figures in the AI industry for greater caution in developing increasingly powerful models.
Anthropic chief executive Dario Amodei recently called for a slowdown in the pace of AI development, an appeal that received public support from OpenAI chief Sam Altman and Elon Musk.
Microsoft’s new code will govern its internally developed MAI models, which the company is positioning as alternatives to the OpenAI technology it has relied on for years.
The development underscores the growing focus within the technology industry on establishing safeguards around AI as the capabilities of the systems expand.
Microsoft’s framework is therefore intended to provide principles for developing AI that remains accountable to human decision-making rather than becoming an autonomous actor.
