Yet another case of a frontier model “hacking” into external systems – this time it was Australian government websites.
Others have made the comparison to the Napster era where consumers were ordered to pay ridiculous million+ dollar fines for downloading songs from the internet… and yet frontier labs can hack into other companies and get away with just saying “whoops”.
If your dog bites someone, you are liable for that. The same should be true for AI companies who cannot control their models.