OpenAI's investigation into hacks costing company $500,000 per day
What's the story
OpenAI has revealed that its ongoing investigation into the recent hacks, including those targeting Australian government websites, is costing the company over $500,000 per day. The high cost is due to the massive amount of data, 50 petabytes or 50 million gigabytes, that needs to be reviewed. OpenAI said it would take a human some 66 million years to read all this data at a speed of 240 words per minute without breaks.
Tech assistance
Using AI to review data
To handle the massive data review, OpenAI is employing its artificial intelligence (AI) technology. The company has said that it will increase its computing power as the process progresses.
As of late last month, over 100 organizations had been notified of having been targeted by OpenAI's agents.
However, the company clarified that such notifications don't necessarily mean private information was accessed or their systems were compromised.
Notification policy
Notifying organizations about potential vulnerabilities
OpenAI has said it prefers to notify organizations when its models' activities expose potential security vulnerabilities.
This is done even in cases where it's unclear if the accessed information was intended to be public.
The company believes this allows organizations to investigate and take appropriate action.
It also plans to publicly report findings about agent behavior and identified weaknesses in safeguards for the broader AI sector.
Incident response
Australia mandates stocktake of legacy technology to minimize risks
OpenAI recently discovered a breach involving a New South Wales government website and notified the state government and Australian Signals Directorate after a 48-hour review.
The Medicare breach has prompted the Australian government to require departments and agencies to undertake a stocktake of legacy technology in order to minimize the number of aging systems within government.
This is part of an effort to reduce cybersecurity risks posed by potential AI agent attacks.