Satya Nadella warned about the dangers of using artificial intelligence services
The rapid development of artificial intelligence technologies is not only opening up new opportunities but also bringing unexpected threats. This was reported by Zamin.uz.
Satya Nadella, head of Microsoft, issued a serious warning to the tech world in his recent statement. According to him, companies using services from major labs like OpenAI and Anthropic may unknowingly be putting the future of their businesses at risk.
As Nadella emphasized, users of AI models today are effectively paying twice. First, they spend money directly on AI services, and second — and most importantly — they provide these systems with their most confidential and valuable business data.
This data is used to further improve the models. As a result, major tech labs learn the inner secrets of their customers and, over time, could become their main competitors.
Many investors and specialists in Silicon Valley compare this risk to swallowing the hook. Companies upload as much data as possible to the systems in hopes of increasing efficiency, and in return, the models learn just as much.
Satya Nadella describes this process as the leakage of intellectual property. In his view, models learn from users' corrections, queries, and usage patterns, absorbing the company's unique experience.
This situation is especially dangerous for new projects and large corporations in developing markets like Uzbekistan. Local companies using global services may inadvertently expose their customer base, strategic plans, and internal workflows.
Nadella defines such knowledge as wealth that competitors can never buy, yet AI models are obtaining it for free. The Microsoft leader also highlighted another important issue.
Companies that create AI models train their systems on openly available internet data for free, but they do not allow others to train on their models. Nadella considers this unfair.
In his view, if labs benefit from open data, then companies should also have the right to create their own smaller and cheaper models based on the results obtained from these models. Currently, some large companies complain that the outputs of their models are being learned by competitors.
However, Nadella calls these restrictions by model creators a joke. He believes that tech giants should also adhere to openness principles and not restrict the benefits of others who learn from their data.
