AI: Microsoft sets out limits on future models, saying ‘it must serve people’

(Il Sole 24 Ore Radiocor) - Microsoft has published a provisional code of conduct that would impose restrictions on its artificial intelligence models, just a few days after the leaders of Anthropic and OpenAI agreed to slow down the pace of development. “We have received feedback from people who want an even more explicit commitment that AI should always work to serve people, not to replace them,” said Mustafa Suleyman, the company’s head of model development, in an interview with CNBC. “We’ve received a lot of feedback that AI should not be addictive or subservient, but should always be there to promote human judgement, autonomy and human action.” Microsoft felt it was appropriate to publish the guidelines now, given recent debates, although it had been working on them for around five months, Suleyman said. Last week, Anthropic researcher Jacob Coxon resigned, stating that the AI lab and OpenAI “are racing headlong towards a self-improving superintelligence and are putting our lives at risk”.

On Saturday, Anthropic’s CEO, Dario Amodei, stated that the pace of improvement in artificial intelligence models needs to be slowed down. OpenAI’s CEO, Sam Altman, expressed his support, and SpaceX’s CEO, Elon Musk, wrote on X that “Dario is right”.

AAA-Pca