OpenAI, the company behind ChatGPT and other generative AI models, has been caught off guard by a Chinese company named DeepSeek, which has developed an open-source model called R1 that achieves similar results to OpenAI’s best-paid models using significantly less computing power.
This unexpected development has shaken investors who have poured money into OpenAI, with the company estimating it won’t turn a profit until the end of the decade. CEO Sam Altman acknowledged that the DeepSeek R1 model is “impressive,” but OpenAI is taking steps to protect its intellectual property.
The company claims that DeepSeek used its models as training data, a practice known as “distillation,” which is against OpenAI’s terms of service. OpenAI spokespersons have warned that they will take countermeasures to protect their technology and collaborate with the US government to safeguard its capabilities.
This incident raises questions about whether using data without permission is a serious issue, especially when it comes to proprietary information. With DeepSeek’s R1 model matching OpenAI’s performance at a fraction of the cost, it remains to be seen how this will impact the US AI industry and its plans for artificial general intelligence (AGI).
Source: https://arstechnica.com/ai/2025/01/i-agree-with-openai-you-shouldnt-use-other-peoples-work-without-permission