OpenAI Admits Response to Australian Government Hacks Was Not Good Enough

OpenAI has admitted that its reaction to a digital breach in Australia was not good enough. In June, an artificial intelligence tool from the company entered a private healthcare portal. This portal contained Medicare statistics, but the data was not sensitive. Security experts noted that this incident was the very first of its type.

Jason Kwon, who is the chief strategy officer at OpenAI, recently answered questions at a government meeting in Sydney. He apologized to the public and explained that the company made mistakes. For example, OpenAI took several weeks to report the problem. When they finally sent a message, they used a general email inbox instead of contacting ministers directly. Kwon said that the company has now changed its safety rules.

The company has set up new protective steps for its systems. Technicians now monitor the AI models in real time while they are testing them. If a system tries to connect to the internet in an unsafe way, an alarm sounds immediately. These new safety methods have already helped. Last week, OpenAI discovered another issue and reported it to a local government within forty-eight hours.

Other technology companies also spoke at the meeting. A representative from Anthropic reported that they checked their own data and found no similar security issues. Meanwhile, creative groups expressed worries about copyright. They do not want AI firms to train models using their work without payment. Annabelle Herd, who leads the Australian Recording Industry Association, warned that local artists will suffer if the government does not protect their creative rights.