- Microsoft CEO Satya Nadella has warned AI companies that train their models on their customers’ trade secrets
- These secrets are then used to train new, more powerful models, which are sold to their customers’ competitors
- But Nadella says there’s a way to stay competitive without being locked into one AI vendor
Microsoft CEO Satya Nadella has warned that the big players in the AI industry are using their proprietary models to learn their customers’ trade secrets, which they can then use to train and deploy more advanced AI models.
The crux of the problem, Nadella said in a blog post, is that “You’re essentially paying for intelligence twice, once with money and again with something even more valuable: the proprietary knowledge you have to reveal to make that intelligence useful. The better you want the model to perform, the more of that knowledge you have to feed it!”
What Nadella is essentially saying is that AI companies harvest sensitive business data from their customers, use it to make training their models cheaper, and then launch those models for use by their own customers’ competitors.
“The kind of knowledge a competitor could never buy”
“Models learn from ‘exhaustion,'” the prompts people write, the tools agents use, and especially the corrections people make when the model is wrong. Each fix is distilled into institutional know-how,” Nadella explained.
Nadella also criticized how AI companies are increasingly complaining about how their models are being distilled by their own competitors. For example, Anthropic accused the retailer and e-commerce company Alibaba of using thousands of Claude calls to distill their own models. By figuring out how a proprietary model works, you don’t have to spend the huge amount of capital needed to retrieve training data and create your own AI model.
This, for Nadella, is a major contradiction in how AI companies operate. “While the great innovation coming from model providers who have fair use rights to train models on public data is necessary, I find it ironic that the status quo is to reverse and impose restrictive terms on distillation,” he said.
It is therefore also hypocritical for AI companies to accuse other companies of distilling their own product and then include clauses in their AI usage contracts that allow AI companies to “reserve the right to learn from customer usage and interaction data.”
“By consuming intelligence, you create intelligence. And what you create should belong to you,” Nadella added.
On-prem is back in fashion
Nadella’s solution to this growing problem? It’s time to go back to on-prem. Nadella encourages companies to “retain ownership” of the data they feed AI models by shifting to the use of “proprietary learning environments” built on the cloud.
The added benefit of moving to these environments is that they allow companies to switch between different AI models provided by different companies using “orchestration layers” and AI gateways.
There is also a growing trend for companies to shift to using open source technologies, which goes hand in hand with companies operating in the cloud. Companies can train open source AI models by using their data already available in cloud environments to do much of what the proprietary models do much more cheaply—and without handing over the same sensitive data that must be used by AI companies to train their own models.
The on-prem solution also has additional benefits. AI models run on-site within manufacturing plants, stores and other premises are far less expensive and require less specialized hardware. Businesses that operate using a centralized cloud increasingly encounter issues with data egress fees, storage bloat and idle specialized hardware.
Google Cloud recently released a report on these very issues and also encouraged companies to move toward using AI gateways and on-prem models to reduce latency, improve resiliency, and reduce cost per cloud. token by switching to local, highly optimized models.
Via TechCrunch
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds.
