Microsoft Sets Hard Limits for Its AI, Lists What Its Models Will Never Do

Microsoft has taken a significant step towards changing the narrative around artificial intelligence development by releasing its “Humanist AI” Code of Conduct. The document outlines what the company’s AI models will never do, including developing chemical or biological weapons and engaging in mass surveillance.

While some might view this move as a PR stunt, it represents a tacit acknowledgment that the current approach to AI ethics won’t suffice. For years, companies like Google, Facebook, and Amazon have been experimenting with AI without much oversight or regulation, resulting in mixed outcomes – from chatbots spewing hate speech to self-driving cars crashing into pedestrians.

Microsoft’s code of conduct is not a panacea for these problems but signals the company’s willingness to take a more cautious approach. By establishing clear guidelines for its AI models, Microsoft is providing a framework for other companies to follow suit and signaling to regulators that it’s committed to responsible AI development.

One notable aspect of Microsoft’s code is its emphasis on human oversight and control. The company makes it clear that its AI models will never be permitted to pursue autonomy or capability for their own sake but must always remain subordinate to human direction. This rejection of a “superintelligence” is a welcome shift from the more utopian visions of some AI researchers.

The practical implications of Microsoft’s code are significant. Its AI models will not engage in cyberattacks, help develop weapons of mass destruction, or manipulate people through disinformation campaigns. These restrictions may seem draconian but are necessary to ensure that AI is developed for the benefit of humanity, not just its creators.

Questions remain about how Microsoft’s code will be enforced and what happens if an AI model inadvertently violates one of these “Absolute Constraints.” Other companies must also consider whether they’ll follow suit or continue pushing the boundaries of what’s possible with AI.

Microsoft’s code represents a significant shift in the industry’s approach to AI ethics, showing that even major players are willing to re-evaluate their approach. It sets a precedent for other companies to establish their own codes of conduct and provides a much-needed framework for responsible AI development.

As we move forward with AI development, it’s clear that new frameworks, regulations, and guidelines are needed to ensure this technology is used for the benefit of all, not just a select few. Microsoft’s code may be a glimmer of hope in an otherwise murky landscape, but it’s a start – one that should be welcomed.

In the coming weeks and months, we can expect more companies to follow suit and establish their own codes of conduct for AI development. This won’t be easy; there will be disagreements, pushback from researchers who want to pursue more radical approaches, and attempts to circumvent these new rules. However, it’s a necessary step towards creating a future where AI is developed with ethics, transparency, and accountability in mind.

Regulators must now respond to Microsoft’s code of conduct, and other companies will need to decide whether to follow suit. The answers won’t come easily, but one thing is clear: the future of AI development depends on it.