Is AI powered IP closer than we thought?

When interviewed by World Trademark Review in 2018, I predicted “No imminent AI apocalypse” and “for anything beyond the most simple, objective trademark-related tasks, he sees little hope that AI could replace an attorney”.

Now I’m not so sure. 

There’s no sign of an apocalypse, but there are clear signs AI is evolving faster than expected.

Take the recent announcements from Open AI. Their Dall-E2 model produces astonishing, original images based only on text instructions. Like the image below in response to the instruction ‘An astronaut riding a horse in a photorealistic style’.




This acceleration is the opposite of what many of us expected. Instead we thought progress would slow, as boundaries within existing hardware and chip design were reached.

Greg Brockman, co-founder of Open AI explained that the stunning improvements come from  the newer model even though it ‘is much smaller & the amount of training compute is similar. Improvements are essentially all due to algorithmic innovation’.

So if improvements are in the software, and not the hardware or training data, then meaningful applications in law and IP may be much closer.

Or how about this joke from Open AI’s GPT-3, when prompted to make a joke involving the dual meaning of regression -

‘Why did the statistician go back to school?
To get his degree in regression!’


So where might the rapid evolution of AI impact on IP and the delivery of IP services?

If AI can construct images and better than average jokes now, it’s not beyond the realms of possibility that within 3 or 4 years it could be
- Processing incoming office actions, setting deadlines and reminders, updating cases and matters, pre-filling standard notifications with filing details
- Summarising matters and documents
- Drafting standard forms from information on matters
- Drafting renewal reports and fee estimates, and check renewal information
- Drafting search reports using results from automated clearance searches, such as we have on LawPanel together with standard paragraphs

And there will be more activities, as registries follow the lead of New Zealand and soon the UK’s IPO with adoption of API first technology stacks.

But there is one caveat. 

Only firms with modern architectures and integrated solutions will have the flexibility and data structures to take advantage of the new opportunities from AI to delight clients and work smarter.

So now is the time to evaluate current systems and processes. And decide whether they’re fit for the future, or will hold your firm back.





Popular posts from this blog

Four of the strangest trademark cases of all time

Big law and the strategic importance of trademark work