Openai stressed the bolts to safety to move away from the eyes of intruders

According to what Openai fixed its security operations to protect against companies spying. According to the Financial TimesThe company accelerated the severity of current safety after the Chinese emerging company Deepseek released a competitive model in January, where Openai claimed that Deepseek copied its models incorrectly using “distillation” techniques.

The enhanced safety includes the “information tent” policies that limit the access of employees to sensitive algorithms and new products. For example, while developing the OPENAI O1, can only be checked members who have been read on the project to discuss the areas of joint offices, according to FT.

There is more. Openai is now isolating royal technology in computer systems that are not connected to the Internet, implementing the biological access to offices areas (wiping employee fingerprints), and maintains an internet policy “rejected-from” that requires clear approval on external communications, for each report, adding that the company has increased material security in data centers and expands its scope on the Internet.

The changes are said to reflect wider concerns about foreign opponents who are trying to steal the intellectual property of Openai, although they are continuing Overall fishing wars Amid American artificial intelligence companies and their increasingly repeated leak Among the comments of CEO Sam German, Openai may try to address internal security issues as well.

We have contacted Openai to comment.