Deepseek surpassed, on Monday (27.jan), the chatgpt in amount of downloads in the Apple app store
Microsoft and OpenAi are investigating a group linked to the Chinese AI (Artificial Intelligence) DeepSek Chinese startup. The goal is to determine whether the data from OpenAi, creator of ChatgPT, were obtained unauthorized. The information was passed by sources anonymous to.
Microsoft security researchers have observed individuals they believe to be linked to DeepSeek allegedly extracting a large amount of data using the application programming interface, the so -called OpenAI API.
Software developers can pay for a license and use the OpenAI API on their products. The type of activity investigated by Microsoft, however, could violate OpenAI’s terms of service or indicate that the group acted to remove artificial intelligence tool restrictions on the amount of data that could be obtained.
OpenAi was notified of the investigation by Microsoft, which is one of the largest investors on the platform.
DeepSeek launched in January a new open source artificial intelligence model called R1. The proposal of the tool is to imitate the way humans reason.
The launch of DeepSeek has made technology actions linked to artificial intelligence – including companies such as Microsoft, Nvidia Corp., Oracle Corp. and Alphabet Inc., Google Controller – designed on Monday (27.jan.2025), with a drop of nearly $ 1 trillion in market value.
David Sacks, “tsar of artificial intelligence ” US President (Republican Party), said on Tuesday (28.Jan) that there “Substantial evidence” that DeepSeek was based on the result of OpenAi models to develop its own technology. In an interview with Fox NewsSacks described a technique called distillation, in which one artificial intelligence model uses another’s results for training purposes to develop similar abilities.
“There is substantial evidence that what Deepseek did here was to distill the knowledge of the OpenAi models. I don’t think OpenAi is very happy with it ”Said Sacks without detailing the evidence.
“As AI leader builders, we endeavored to protect our intellectual property, including a careful process to include border abilities in the launched models, and we believe, as we advance, that it is extremely important that we work closely with collaboration with The US government to better protect the models that are able to efforts from opponents and competitors to take US technology ”said an OpenAi spokesman in a statement.
