Of all the debates raging about the potential downsides of AI, there is one worry causing the most hand-wringing among AI enthusiasts in Silicon Valley — that the giant AI labs that sell proprietary models are somehow acting like Trojan horses.
quoting
naddr1qv…ups8Microsoft CEO Satya Nadella has warned that companies using proprietary AI models are inadvertently paying twice by exposing sensitive business data that trains these models. He advocates for companies to retain data ownership by building their own cloud-based learning environments and using orchestration layers to switch between AI providers. This aligns with a growing trend of enterprises adopting on-premise, open-source AI models to gain control and reduce costs. - Companies using proprietary AI models from labs like OpenAI and Anthropic risk exposing sensitive business data, which the model providers can use to their advantage. - Microsoft CEO Satya Nadella states users pay twice: once with money for AI usage, and again by revealing proprietary knowledge to train the models. - Enterprises are teaching AI models about their businesses through prompts, tool usage, and corrections, essentially handing over valuable institutional know-how. - Nadella argues that if AI companies can train on public data, users should be allowed to 'distill' their models in return. - He urges companies to retain data ownership by building proprietary learning environments on the cloud and using orchestration layers to avoid vendor lock-in. - There is a growing trend towards enterprises adopting open-source AI models on-premise, seeking cost savings and greater control over their data. - Companies like Vercel and OpenRouter are seeing increased traffic to open-source models, indicating a shift in enterprise AI adoption. Continue reading https://techcrunch.com/2026/07/13/satya-nadella-has-issued-a-shocking-warning-to-companies-using-ai/