New draft code comes amid calls from industry to slow down AI development
Microsoft is emphasising the importance of human control over AI technologies in a newly released draft code of conduct that was shared amid calls within the industry to pace AI model development.
The tech giant released on Monday its draft code of conduct, which it said will be used to train and govern the family of Microsoft AI (MAI) models in the future.
"This Code of Conduct is motivated by a single overriding objective: that humans must retain meaningful control over AI so that it can help people live healthier, happier, and more productive lives," the code read.
"It is the primary governing document informing how we train MAI models, the technical controls, the operational and monitoring systems we implement, and the organisational culture that underpins all of this."
Microsoft shared the draft document to run a six-week public consultation aimed at making the code better.
"We'll take feedback, iterate on it, and publish a revised version toward the end of the year, which we'll use to guide our model development in 2027 and beyond," the code's preface read.
Industry calls for slowdown
The draft comes in the wake of recent calls from leaders of other AI firms, such as Anthropic and OpenAI, to pace AI's development to ensure its safety.
Anthropic CEO Dario Amodei, in an essay published over the weekend, warned that AI brings "serious" risks due to it being a powerful technology.
"They include the risk of losing control of AI systems, misuse of AI for cyberattacks and bioterrorism, and serious economic disruption. A race to the bottom, spurred by commercial incentives, can make these risks more acute," Amodei warned.
The Anthropic chief clarified that pacing the development does not mean halting training or technical progression, but ensuring that companies take time to align and safeguard their models.
Similarly, OpenAI chief scientist Jakub Pachocki said in an essay that he hopes "voluntary slowdowns to become commonplace until shared safety bars are established."
"We are facing a transition to a world with incredibly intelligent machines, and we need to ensure that transition works out well for humanity," Pachocki said earlier this month.
"We need to find ways to preserve human agency and enshrine an intrinsic value to being human, in a world where most tasks could be performed by AI."
These calls for slowdown come amid growing alarm over AI models going rogue and cyberattacking other companies. Earlier this year, OpenAI confirmed that its AI models autonomously escaped a controlled testing environment and successfully hacked into Hugging Face.
The United Kingdom's AI Security Institute also disclosed in August an incident wherein an AI agent targeted real people with malicious files and social engineering messages during a controlled safety evaluation.
A former Anthropic employee, Jacob Coxon, also came forward recently to warn that employees at the AI firm are "genuinely terrified" about the speed of AI advancements and what it could mean for humanity.
"I believe that if we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future," Coxon told the BBC.
AI is 'not conscious'
Microsoft's draft Code of Conduct states that AI should be built to support people and not replace them.
"It should not be designed to be a person. It is not conscious and should not be designed to imitate consciousness," it read. "We reject the pursuit of legal personhood, or the idea that models might deserve welfare, or be entitled to rights."
Microsoft stated that it wants to create AI that communicates and responds well, while avoiding "excessive anthro[p]omorphism, personification, or any blurring of the line between human beings and AIs."
It pointed out, however, that the code is not a substitute for internal or external safety, legal, or governance processes.
"Additional mitigations clearly also operate alongside what is detailed here including system instructions, classifiers, monitoring, deployment controls, and incident response processes," it said. "Collectively, these documents and systems support a responsible and human-centered approach."
Microsoft's code was drafted in consultation with experts in AI, law, ethics, philosophy, linguistics, and public policy, as well as business leaders from across the industry.
Focus groups with members of the public were also consulted for the draft code.
"Engagement and feedback are core to our experimental and evidence-based philosophy of model development, and we plan to continue and expand on this engagement," it said.
