OpenAI tightens the screws on security to keep away prying eyes

It is reported that the Open has overcomes its security activities to protect against corporate spying. According to the Financial TimesThe company has accelerated an existing security clampdown after the Chinese Startup DIPSEC released a competitive model in January, OPNA alleged that DEPSEC had incorrectly copied its models using “Patan” strategies.

Biffid-up protection includes “data tents” principles that limit the access to sensitive algorithms and employees on new products. For example, during the development of the Openai and 1 model, only the members of the verified team who read in this project can discuss it in the shared office places, according to FT.

And there are more. OpenAI is now disintegrating the technology owned on the offline computer system, applying biometric access control (it scans the staff fingerprints) and a “denying” Internet policy for external connections, which further adds that the data of the agency has increased and its physical protection has increased and its syllabi.

These changes are asked to reflect the extensive concern about foreign opponents that try to steal the Buddhist property of Openai, although ongoing have been given Pooching fight Between American AI agencies and the growing frequently Leaked In the comments of CEO Sam Altman, Open can also try to solve internal security problems.

We reached the opening to comment.