Microsoft AI Code of Conduct Sets New Rules for AI Safety, Human Oversight and Responsible Artificial Intelligence Development

Microsoft has released a new AI code of conduct designed to guide how its AI models are trained and used. The document focuses on AI safety, AI alignment and human control, as concerns about increasingly powerful AI systems continue to grow.

Microsoft said that superintelligent AI could eventually perform better than humans across many tasks. Because of this, the company believes it is important to clearly define how such systems should be developed and controlled.

The new Microsoft AI code of conduct outlines basic principles that AI models should follow. One key idea is that AI should support people rather than replace them and should be developed in ways that benefit humans. The document also includes strict safety rules. Microsoft's AI models are not allowed to support activities such as cyberattacks, nuclear weapons or the creation of deepfakes.

The rules also focus on keeping humans in control of AI systems. Microsoft says its models must not use deceptive or self-reinforcing methods to avoid human oversight. According to the company, AI models should not try to prevent authorised people or systems from directing, changing or shutting them down. This is an important part of Microsoft's approach to AI safety and alignment.

Microsoft's announcement comes at a time when AI safety has become a major topic across the technology industry. AI companies are facing increasing questions about how powerful models should be tested, controlled and monitored.

Microsoft is among several major AI companies working on ways to make advanced AI systems safer. The company has also shown support for ideas such as embedded evaluators, which can help monitor AI systems during their development and use. Microsoft CEO Satya Nadella stated that the company supports research and careful development aimed at getting AI alignment right.

The new Microsoft AI safety rules give a clearer picture of how the company wants its AI models to behave as the technology becomes more powerful. For users, the focus is on keeping AI systems useful while ensuring that people remain responsible for important decisions and control over the technology.

This article is based on information from Tech Crunch