- Microsoft CEO Satya Nadella has warned that AI companies are training 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 is a way to stay competitive without being limited to a single AI vendor.
Microsoft CEO Satya Nadella has warned that 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 twice for intelligence, once with money, and again with something even more valuable: the proprietary knowledge you have to reveal to make that information useful. The better you want the model to work, the more you have to feed it with that knowledge!”
What Nadella is essentially saying is that AI companies collect sensitive business data from their customers, use it to make training their models less expensive, and then release those models for use by their own customers’ competition.
“The kind of knowledge a competitor could never buy”
“Models learn from ‘exhaustion,’ from the prompts people write, from the tools agents use, and especially from the corrections people make when the model is wrong. Each correction 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 retailer and e-commerce company Alibaba of using thousands of Claude prompts to distill their own models. By understanding how a proprietary model works, you don’t need to spend the huge amount of capital required to obtain training data and create your own AI model.
For Nadella, this is a major contradiction in the way AI companies operate. “While the great innovation brought by model providers with fair use rights to train models on public data is necessary, I find it ironic that the status quo must then turn around and impose restrictive conditions on distillation,” he said.
So it’s also hypocritical for AI companies to accuse other companies of distilling their own product and then including clauses in their AI user 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 be yours,” Nadella added.
On-site is back in fashion
Nadella’s solution to this growing problem? It’s time to get back to on-site. Nadella encourages companies to “retain ownership” of the data they feed into AI models by moving 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 of businesses moving towards open source technologies, which goes hand in hand with businesses operating in the cloud. Companies can train open source AI models using their data already available in cloud environments to do much of what proprietary models do, for much less – and without passing that same sensitive data to AI companies to train their own models.
The on-premises solution also has additional benefits. AI models operated on-site in manufacturing plants, stores and other premises are much cheaper and require less specialized hardware. Businesses that operate using a centralized cloud increasingly face issues with data egress charges, excessive storage, and idle specialist hardware.
Google Cloud recently released a report addressing these same issues and also encouraged businesses to use AI gateways and on-premises models to reduce latency, improve resiliency, and reduce costs per token by moving to highly optimized on-premises models.
Via TechCrunch
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.
