OpenAI, the organization responsible for the popular ChatGPT AI, is voicing discontent over DeepSeek, a new Chinese AI assistant it accuses of utilizing OpenAI’s technology without authorization. The appearance of DeepSeek, which reportedly was developed at a significantly lower expense than other AI models yet claims similar capabilities, has negatively impacted stock prices for several companies focused on AI, such as Nvidia.
A recent report from Bloomberg highlights that Microsoft, an investor in OpenAI, is probing whether DeepSeek has appropriated a substantial amount of data from OpenAI. OpenAI’s licensing permits developers to access its API to integrate it into their programs, implying that DeepSeek may have trained its capabilities using OpenAI’s responses.
*“There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models and I don’t think OpenAI is very happy about this,” stated David Sacks, the new AI czar. *Translation: “There’s a lot of evidence showing that DeepSeek extracted knowledge from OpenAI models, and I believe OpenAI is not pleased about it.”
Distillation refers to when an AI model utilizes outputs from another model to enhance its training. OpenAI has also commented on this issue, asserting: *“Companies based in the Peoples Republic of China and others are continually attempting to distill models from leading US AI firms. As the foremost creator of AI, we are implementing measures to safeguard our IP… and deem it essential to collaborate with the US government to defend our top-tier models against potential threats from adversaries and competitors.” *Translation: “Companies, particularly from China, are trying to replicate US AI models. We are ensuring our technology is protected closely with the government.”
Prior discussions in the industry indicate contrasting views on the ethical aspects of AI training. Crystal van Oosterom from OpenOcean noted, *“DeepSeek has clearly built upon publicly available research from major American and European institutions and companies.” *Translation: “It’s evident that DeepSeek has utilized publicly accessible research from prominent US and European firms.”
Moreover, OpenAI is facing allegations of having scraped extensive internet resources indiscriminately for training its models. With ongoing litigation from The New York Times against OpenAI and Microsoft, more publishers globally are increasingly challenging the methods employed by OpenAI. Ed Zitron, a tech critic, voiced his amusement, remarking, *“OpenAI, the company built on stealing literally the entire internet, is crying because DeepSeek may have trained on the outputs from ChatGPT. They’re crying their eyes out. What a bunch of hypocritical little babies.” *Translation: “OpenAI, which has taken data from the internet, is upset that DeepSeek might have learned from their outputs. They seem very contradictory.”
In an anecdotal exchange, the DeepSeek chatbot claimed, *“No, I am an intelligent assistant developed by the Chinese company DeepSeek, built on our own proprietary technology and learning models. We respect intellectual property rights and adhere to stringent ethical standards in developing our AI protocols.” *Translation: “I’m a chatbot created by DeepSeek with our unique technology. We honor IP rights and follow strict ethical guidelines.”
