- Microsoft CEO Satya Nadella warned that artificial intelligence companies are training their models using their customers’ trade secrets.
- These secrets are then used to train new, more powerful models that are sold to your clients’ competitors.
- But Nadella says there is a way to stay competitive without being tied to a single AI vendor.
Microsoft CEO Satya Nadella 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 fact of the matter, Nadella said in a blog post, is that “you basically pay twice for intelligence, once with money and once with something even more valuable: the unique knowledge that must be revealed for that intelligence to be useful. The better you want the model to work, the more knowledge you have to feed it!”
What Nadella is essentially saying is that AI companies are collecting sensitive business data from their clients, using it to make it cheaper to train their models, and then releasing these models for use by their own clients’ competitors.
“The kind of knowledge a competitor could never buy”
“Models learn from ‘exhaustion,’ from the prompts people type, from the tools agents use, and especially from the corrections people make when the model is incorrect. Each correction is distilled into institutional knowledge,” Nadella explained.
Nadella also criticized how AI companies are increasingly complaining about how their own competitors are distilling their models. For example, Anthropic accused the retailer and e-commerce company Alibaba of using thousands of Claude’s cues to distill their own models. By figuring out how a proprietary model works, you don’t need to spend the enormous amount of capital required to obtain training data and create your own AI model.
This, for Nadella, is a major contradiction in the functioning of AI companies. “While the great innovation that comes from model providers having fair use rights to train models with public data is needed, I find it ironic that the status quo then turns around and imposes restrictive terms on distillation,” he said.
Therefore, it is also hypocritical for AI companies to accuse other companies of distilling their own product and then include in their AI usage contracts clauses that allow AI companies to “reserve the right to learn from customer interaction and usage data.”
“By consuming intelligence, you are creating intelligence. And what you create should belong to you,” Nadella added.
The place is back in fashion
Nadella’s solution to this growing problem? It’s time to return to the local environment. Nadella encourages companies to “retain ownership” of the data that feeds AI models by shifting to using “proprietary learning environments” built in the cloud.
The additional benefit of migrating 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 move to open source technologies, which goes hand in hand with companies 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, much cheaper, and without handing over that same sensitive data for AI companies to use to train their own models.
The on-premise solution also has additional benefits. AI models operated on-site within manufacturing plants, stores and other facilities are much cheaper and require less specialized hardware. Businesses operating using a centralized cloud are increasingly facing issues with data egress fees, excess storage, and down specialized hardware.
Google Cloud recently published a report on these same issues and also encouraged enterprises to move toward using AI gateways and on-premises models to reduce latency, improve resiliency, and reduce costs per token by switching to highly optimized on-premises models.
Through TechCrunch
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds.
