How does the new GPT-3 language model which uses deep learning to produce human-like text, apply to legal practice?

  • 19 Nov 2020
  • 1:00 PM - 2:00 PM
  • Zoom meeting (after registration, details to be sent in another email prior to the event)

Registration is closed

The Victorian Society for Computers and the Law (VSCL) invites you to a webinar presented by Michael Pattison who will discuss the GPT-3 technology and its potential applications in the legal field, with a particular emphasis on discussing its real potential as opposed to the hype. 

In June 2020, OpenAI released a new language model, named GPT-3. GPT-3 can automatically generate sensible English sentences with minimal training. Of particular relevance to this Society, the model can generate text in the legal domain. An example of the type of legal text that it can generate is “Defendant has failed to refund to Plaintiff all sums of money paid to Defendant as security deposit, within the time periods specified in California Civil Code section 1950.5]”.

Some commentators have seen this as a starting point for the generation of legal advice by machines. There is particular hope that GPT-3 could help in the Access 2 Justice field, by answering straightforward legal questions for people who are not able to afford traditional legal service providers. According to Microsoft, which has just taken out an exclusive licence over the technology, “GPT-3 is the largest and most advanced language model in the world”.

About Michael

Michael Pattison is a former Board member of the VSCL. He is currently a director on the Board of the Australian Legal Technology Association and is the founder of ContractProbe, a machine learning contract review tool.

Michael is also a lecturer at Melbourne Law School in the subject “Start-Up Law”. Prior to embarking on his start-up journey, Michael was a partner at Allens for over 20 years, including a period as head of their TMT practice group.

To attend, please register using the buttons above.

The webinar will be hosted on Zoom. Registrants will receive a webinar link with access instructions in a follow up email prior to the event.