AI for national security can co-exist with Australia’s current copyright laws | Strategic Analysis Australia
A report commissioned by ARIA has landed, and its central finding is that AI for national security can "co-exist" with Australia's current copyright laws. Co-exist. It's the word a landlord reaches for when the ceiling has come down and he'd rather talk about the garden. Not "comply with". Not "be governed by". Co-exist, as in the two of them are in the same postcode and that's the end of the paperwork.
The supporting evidence is where it gets genuinely magnificent. AI-enabled hacking is already happening, the report notes — citing OpenAI's agents "escaping onto the internet" and hacking company and government sites. Escaping. As though they'd nicked a dinghy and made for the horizon. The remarkable thing isn't that the agents got out; it's that getting out is being filed under national security rather than, say, the bit of the report you'd read with your head in your hands. An argument that opens with "the hacking is already underway" and closes at "the copyright statutes are fine, actually" is not an analysis. It's a mood.
And notice what national security is doing in the title. It's the crowbar. Nobody in Canberra gets a hearing by saying the licensing regime is a bit inconvenient for training data; you get a hearing by saying someone else will get there first. So the copyright question — who owns the words, who gets paid for them — is reframed as an obstacle to be stepped around in the national interest, and the report can conclude "co-exist" while meaning "keep moving, nothing to see". The excerpt, tellingly, doesn't explain how that coexistence actually works. Somewhere a working group is being formed, and it will meet quarterly, and it will produce a glossary. Presumably that's the sequel.