In just a couple of days, Depseek has managed to put the artificial intelligence industry up, presenting a new model, called R1, which is capable of doing the same as the best chatgpt model so far, but completely free.
Using Deepseek is very easy, both in mobiles and on the web, and only requires the creation of a free account; As if that were not enough, use a model ‘Open Source’ that releases the code so that anyone can use it in their own projects. It has been a ‘coup on the table’ that has awakened investors and experts of his idyll with AI, and many expected the response of the big US companies.
Although the initial reaction of leaders such as Sam Altman of OpenAi against Deepseek was diplomatic, praising his technical achievements, he has not needed to wait long for the first accusations of dirty game to arrive; In fact, there are already some voices that They ask for the Deepseek prohibition In the United States, officially due to intellectual property robbery but in practice, to protect its technological industry.
The first accusations have arrived precisely from Openai, who has assured Financial Times that has evidence that Depseek has used chatgpt to train and improve its artificial intelligencetaking advantage of the work of their engineers. Specifically, the company claims to have found ‘distillation’ tests, an automatic learning process based on another AI; In other words, he would mean that R1 of Deepseek would have been trained using the responses generated by Chatgpt, analyzing their behavior and replicating it.
Actually, ‘distillation’ is a very common process in the development of generative, to transfer the advances from one AI to a new one; However, the terms of service clearly They prohibit using chatgpt to ‘distilled’ another AIprecisely to prevent their rivals from simply copying the advances obtained by Openai.
This is a serious accusation, which aims to show Deepseek and its R1 development process; The message is clear: Openai does not believe that Depseek has been able to develop an AI capable of competing with chatgpt with a budget fractionas stated.
For its part, Microsoft, Openai’s great ally, has been somewhat more cautious, reporting Bloomberg which is “investigating” the possible use of chatgpt from Deepseek. This complaint adds to that already made OpenAi employees that Depseek get too many personal data from userswhich can then sell to ‘commercial partners’.
The free android
Of course, nobody escapes the irony that now OpenAi is very concerned about intellectual property and data theft. We talked about the same Openai that was ‘caught’ allegedly training its AI with YouTube videos without permission, which was denounced for stealing personal data to train its AI, and which was notified by violating the EU Data Protection Law .
The question is now what will happen to Depseek after these accusations. The head of the Government of Donald Trump, David Sacks, has confirmed Fox News that there are “substantial tests” of the ‘distillate’ of Chatgpt from Deepseek, without giving more details. In the worst case, Depseek could follow the steps of Tiktok and be prohibited in the United States.
