All articles
Yesterday's Senate AI inquiry in Sydney, Google said that if developers had to get "prior…
Yesterday's Senate AI inquiry in Sydney, Google said that if developers had to get "prior authorisation from everyone", then "I just can't see how AI development can happen."
Lawyers should read that as a due diligence question, not a policy debate.
At the same hearing, the ABC said Australia's copyright system "is completely adequate to manage the business of AI".
Canberra has already leaned that way. In October 2025 it ruled out a text and data mining exception. And last week a US appeals court held that training a legal research tool on Westlaw headnotes was not fair use.
So in Australia, the default is permission. That matters to us twice.
First, the tools we buy. If a vendor can't tell you what its model was trained on, you're carrying a risk you can't explain to your client.
Second, the data we hold. Our precedents, advice and client documents are copyright works and confidential information. Our duty of confidentiality doesn't get rewritten by a vendor's terms of service.
Before your next AI renewal, ask two things in writing: what was this trained on, and will our data train it?
When did you last read the training clause in your AI vendor's terms?
#AusLaw #LegalTech #AICopyright
View on LinkedIn