Artificial Intelligence in Conflict: Microsoft investigates whether Deppseek stole OpenAi data

Artificial Intelligence in Conflict: Microsoft investigates whether Deppseek stole OpenAi data

The iDeepseek’s rruciothe company of Artificial Intelligence from Chinacaused a stir since his first appearance, both for the News that presents, as for its efficiency. Is that it is a model of Open source R1 with cheaper and less advanced chipswhich questioned the high expense that American companies They perform in this technology.

The conditions that evidenced made their noise in the market stronger, and that another technological giant such as Microsoft opens an investigation in order to discover if there was a Obtaining data in an unauthorized way of OpenAI, owner of chatgpt.

As indicated, Security researchers of the company Bill Gates They observed how individuals who believe can be related to Deepseek, They leaked a lot of data using the Application programming interface (API) of the company led by Sam Altman.

In this sense, it should be noted that Software developers They can pay a license to use the API and integrate OpenAI patented artificial intelligence models into their own applications.

Microsoft suspicions

Microsoft is Openai’s technological partner and its largest investor, so it notified the company about this suspicious activity, which could violate the terms of service of the Chatgpt creator.

In addition, as they point out, it could also indicate that The group acted to eliminate OpenAI restrictions about the amount of data they could obtain.

The US intervention

On the other hand, for David Sacks, the person in charge of AI in the US government, Donald Trump, there is one “Substantial evidence that Depseek was based on the results of Openai models to develop your own technology. “

David Sacks.jpg

David Sacks talked about the problems with Deepseek.

CNBC

During an interview, Sacks described a technique called distillation, through which an AI model uses the results of another for training purposes to develop similar capacities.

“There is substantial evidence that Deepseek made here was Extract knowledge of OpenAI models And I don’t think the company is very happy with this, “Sacks said.

What did they say from OpenAI

On the other hand, OpenAI, and in response to these comments, said: “We know that companies based in the China People’s Republic and grant countries, They are constantly trying to distill the models of the main artificial intelligence companies in the United States. “

“As the main constructor of artificial intelligence, We take countermeasures to protect our intellectual propertyincluding a careful process to determine which avant -garde capabilities include in published models, “he continued

“We believe that, as we advance, It is of vital importance that we work in close collaboration With the United States government for better protect the most capable models of the efforts of adversaries and competitors For seizing American technology, “he concluded.

Source: Ambito

Leave a Reply

Your email address will not be published. Required fields are marked *

Latest Posts