Microsoft is reportedly probing into allegations against DeepSeek over inappropriate usage of OpenAI’s APIs. OpenAI’s models were supposedly used by DeepSeek for training their own AI systems, as per claims made by David Sacks.
It is believed that DeepSeek, backed by a Chinese firm, might have carried out substantial data extraction using OpenAI’s API during autumn 2024. As the largest shareholder of OpenAI, Microsoft intervened upon detecting dubious activity, alerting OpenAI accordingly.
According to OpenAI’s terms of service, while its API is accessible to all, using the output to build a new AI model is strictly not allowed. OpenAI explicitly forbids the development of competing models using its output.
The bone of contention essentially lies in the method of distillation employed by AI model developers to gather knowledge from another model. It remains to be seen if DeepSeek has devised imaginative ways to bypass OpenAI’s rate limits and extensively query its API. Should this hold true, it assuredly invites legal consequences.
Original source: Read the full article on TechCrunch
