DWA Law
For the past year or so AI has been part of almost every working day at DWA. We have used it for legal research, proofreading and first drafts of agreements. We have received contracts back from the other side that were very clearly marked up by a machine. And producers and creatives have started asking us questions that nobody was asking a year ago. Who owns a script if an AI tool helped write it? Can I use AI concept art in my pitch deck? What am I agreeing to if I let a studio scan my face?
Those questions are now a regular part of our practice, so we are making AI a central focus of how we work and what we advise on. That focus is for the people we have always acted for, including producers, writers, directors, performers and the companies making screen content in Australia.
A year of AI on the other side of the table
Most lawyers will recognise the scenario. A contract goes out for review and comes back with a markup that bears all the hallmarks of a machine. Every typo has been corrected. The clause numbering is flawless. It even picks up where the same word has been used twice in a sentence.
Then come the problems. Sweeping arguments that have little to do with the parties or the commercial deal. Requests for caps on indemnities that no one in the market would expect. Firmly held positions with no reasoning behind them. There are other giveaways too, such as an unusual number of em dashes and a turnaround time that is suspiciously quick.
One of DWA’s lawyers was once told by a university torts lecturer that studying law was a mistake because AI would make lawyers irrelevant. It is a prediction we have thought about a great deal this year. Our view, so far, is not yet, and not in the way many people expect.
Where AI genuinely helps
The time saving is significant. We have prepared a first markup of a document in an hour that would ordinarily take a couple of days. But that first markup is not legal advice. The advice comes afterwards, when a lawyer reviews the document carefully, and that step cannot be skipped.
AI is also very good at the detail that matters so much in contract work, such as inconsistencies, defined terms that do not match and the typo that changes the meaning of a clause. When so much of a negotiation turns on how a document will be read, that is genuinely useful.
For producers and creatives it can also be a helpful way to understand a contract. Asking an AI tool to explain a deal in plain English before you sign is sensible, and it means you come to the conversation with your lawyer better informed. The caveat is to think carefully about what you upload and where it goes, because publicly available AI tools generally do not keep information confidential.
Where AI falls short
AI is not actually intelligent, and what it produces depends heavily on what it is given. The best way we can describe working with it is this. Imagine standing in front of the Mona Lisa next to an artist friend who is blindfolded. Your task is to talk them through painting a perfect copy, except with Homer Simpson’s face. You tell them where to paint, which brush to use, which colour and how warm the light should be. If you are very good at it you might get them 60% of the way there. But some things are too difficult or too slow to describe, and you cannot describe what you do not know you need to mention.
Deals are like that. A film deal is not just a number of parties and a few pieces of legislation. It is the client’s appetite and budget for a dispute, the relationship between a producer and a writer who want to work together again, and all of the context that never makes it into writing.
We have seen those limits first hand. When we tested an AI platform on a complex matter, carefully following the prompting guidance, the volume of detail overwhelmed it. Its answers began to contradict each other, and at one point it introduced a person’s name that appeared nowhere in the material. At some point it crossed the line from useful to unreliable, and it gave no indication that it had done so.
The courts have seen the same problem. In August 2025 a senior barrister apologised to the Supreme Court of Victoria after submissions in a murder trial were found to contain fabricated quotes and case citations generated by AI. In 2026 the Court of Appeal referred a solicitor to the Legal Services Commissioner in Oberoi v Douglas [2026] VSCA 31 after finding cases in the material that did not exist.
Lawyers therefore need to be just as careful as anyone else. The regulators’ joint statement on the use of AI in Australian legal practice, the Law Society of NSW’s guide to the responsible use of AI and the Federal Court’s Generative AI Practice Note all set out what is expected of practitioners who use these tools, from protecting client confidentiality to verifying every output before it is relied on.
The part that concerns us most
Our greatest concern is the party on the other side who is relying on AI in place of a lawyer. In a cooperative contract negotiation, where everyone wants the deal done, that can work for a while. But once there is a genuine back and forth it becomes apparent, and that party is left relying on the other side’s lawyers not to take advantage.
That raises an ethical question the profession has not yet answered. The conduct rules address how a solicitor should deal with an unrepresented party. They do not yet address a party who is, in effect, represented by a chatbot.
In a dispute or settlement negotiation the stakes are higher again. Once the other side can see that a party’s responses are AI-generated, that party has lost ground, and someone with a genuine claim may walk away with nothing. If you have suffered a real loss and believe you are entitled to something, we would strongly encourage you not to rely on an AI tool to represent you.
Where the law is heading
Parliament is examining these questions now. The Joint Select Committee on Artificial Intelligence was established in August 2026 and is due to deliver its final report by 30 November 2026. Its terms of reference are broad. In summary the committee is considering the opportunities, risks and harms of AI, its effect on jobs and skills, national security and data sovereignty, and whether Australia’s existing laws (including copyright) remain fit for purpose, along with any other relevant matter it identifies.
Submissions closed on 14 September 2026 and the published submissions are well worth reading. Creative industry bodies such as Screen Producers Australia and the Media, Entertainment & Arts Alliance called for consent, fair payment and transparency when creative work is used to train AI, and for stronger protections for creative workers. Global technology companies and some policy groups argued for copyright settings that would make it easier to train AI in Australia. Universities and research centres focused on accountability, privacy and automated decision-making, and AI safety organisations called for stronger oversight of the most powerful models.
Read together, the submissions point to several gaps in the current law. There is no settled position on how copyright applies to AI training or to AI-generated material, although Australian courts have held that copyright requires a human author. Performers have limited protection against digital replicas of their face or voice, because Australia has no general image or personality right. The Privacy Act 1988 (Cth) still exempts most small businesses and does not squarely address technologies such as facial recognition. And when AI causes harm it is often unclear who in the chain from developer to user is responsible. We will be following the committee’s report and recommendations closely.
Why we are leaning in
The screen industry already has. AI is appearing in the script, the edit, the VFX pipeline, the dub and the marketing, and in the contracts that surround all of it. The legal questions are arriving faster than the law is changing. Distributors, insurers and financiers are starting to ask which tools were used and who owns what they produced, and funding bodies are moving the same way. Screen Australia, for example, lists transparency among its AI guiding principles.
Our view after a year of working with AI is fairly simple. It can match most lawyers on administrative and word processing tasks. We have yet to see it give advice that would compete with a lawyer a few years into practice. But it can make every lawyer more efficient, whether they have been practising for two years or twenty, provided a person remains responsible for the advice.
That is how we use AI at DWA, and it is how we will help our clients use it. As a tool, with human judgement at every step that matters.
If there is an AI question you would like us to explore, please leave it in the comments or get in touch. You can contact David Whealy at david.whealy@dwalaw.com.au or Brooke Clarke at brooke.clarke@dwalaw.com.au.
Related article
This article is general information only and is not legal advice. Speak to us about your specific circumstances.