Microsoft investigates the legality of DeepSeek’s R1 model

Microsoft is investigating the Chinese company DeepSeek for allegedly using nefarious methods to train its AI models, potentially violating Microsoft's terms of service by using OpenAI's API. DeepSeek's R1 model, which has gained attention for its cost-effective training, may have been developed using a process called distillation, where one model learns from another. Microsoft suspects that DeepSeek extracted code from OpenAI's API, raising concerns about intellectual property theft. OpenAI emphasizes the importance of protecting its models from international competitors and is working with the U.S. government to safeguard its technology.

In other news, Microsoft is revising its Recall feature in response to backlash, making it opt-in rather than opt-out on Copilot+ PCs. Meanwhile, Apple is reportedly developing its own large language model, ReALM, to enhance Siri's capabilities, potentially rivaling OpenAI's ChatGPT. Additionally, Microsoft’s Copilot, a competitor to ChatGPT, is available for integration into various Microsoft tools, offering users a versatile AI chatbot experience.

Key takeaways

Microsoft is investigating DeepSeek for potentially violating terms of service by using OpenAI's API to train its AI models.
DeepSeek is praised for its cost-effective AI model training, but there are concerns about its methods, including possible distillation from OpenAI's models.
Microsoft has made changes to its Recall feature in response to backlash, making it opt-in instead of opt-out on Copilot+ PCs.
Apple is reportedly developing its own large language model, ReALM, to enhance Siri's capabilities, potentially rivaling OpenAI's ChatGPT.

Microsoft investigates the legality of DeepSeek’s R1 model

Key takeaways

Discussion (0)