DeepSek Accused of Using Gemini Training Data
A company known as DeepSek AI is currently under close watch. Reports suggest the company may have used output from Google’s Gemini AI model to train its own artificial intelligence system. This method, called distillation, involves using a powerful AI model to teach a new one, which is seen as an efficient way to train models.
However, this practice raises serious questions about ethics. Some major AI companies have policies that forbid using their output to train competing models. DeepSek has faced similar claims in the past concerning the ChatGPT model.
DeepSek gained attention earlier this year with an AI model said to rival those from larger companies. Experts in technology are now suspicious that DeepSek’s latest update used Google Gemini as a direct source for training data, not just general information.
These suspicions grew after observations were shared online about the way DeepSek’s new model responds and processes information, noting similarities to Gemini.
This is not the first time DeepSek has faced such accusations. When it first launched, there were claims it used output from OpenAI’s ChatGPT in its training. This was suggested as a reason why DeepSek’s training costs were lower than competitors.
The distillation technique DeepSek is thought to use is not a new concept. It is often described like a teacher passing simplified knowledge to a student. Using a highly developed AI model to train a new one can be very efficient in terms of cost and time.
But this efficiency comes with risks of breaking rules and ethical standards. If it is true that DeepSek used the output of other models against their terms, it would be a violation of those policies.
Despite the ethical concerns, some experts view DeepSek’s approach as understandable from a business point of view. It is suggested that a company with limited access to computing power but available funds might use this method to effectively gain more training capability.
Geopolitical issues, including limits on access to advanced technology like computer chips, can create challenges for companies developing AI. Facing these pressures and needing to compete globally, some companies might choose practical paths, even if they involve ethical or legal risks.
The situation with DeepSek highlights a lack of clear rules in the rapidly advancing AI industry. It shows a conflict between using efficient methods and following ethical principles and respecting intellectual property rights. How such practices will be handled in the future of global AI remains uncertain.
