Generative AI models, in particular – including large language models (LLM) – could be enhanced in the near future so that they are tailored to work optimally for in-house counsel and legal use cases, he said.

These models could be used to accelerate the work in-house counsel are currently doing, including drafting emails and responses, briefings to external counsel, and researching the application of and changes to the law. They can also detect similarities and differences between different documents, which could prove useful when comparing old and new contracts and identifying significant changes.

His comments preceded the summit, where he and a panel of speakers will explore how in-house counsel could harness generative AI models to propel their businesses and outline the features the next generation of LLMs could offer.

Dickason recently wrote in Lawyers Weekly that these models are currently trained on large datasets containing vast amounts of material that is not relevant to legal use cases. However, fine-tuning LLMs for specific legal tasks could help the model understand domain-specific legal phrasing and terminology.

He told Lawyers Weekly that once LLMs are customised for legal use, they could change the structure of the legal industry, including the types of tasks carried out by in-house counsel.

“This means in-house counsel would have the opportunity to in-source and do more exciting work internally, meaning they don’t have to go to external panels for certain types of advice,” he proposed.

For example, he explained, while in-house counsel may currently consult their external panel for advice on employment law, they could insource some of these tasks with assistance from the appropriate research platforms, as well as LLMs and other generative AI tools.

VIEW ALL

Dickason likened this to the automation of processes in conveyancing and writing wills, which has enabled legally trained professionals and non-lawyers to execute these tasks.

“This is going to force legal work to move up the value chain. We’re going to see the next level of legal advice or legal workflows become automated as a result of processes like this,” he said.

As a result, law firms on external panels would have to demonstrate their value by providing services like horizon scanning to identify signals of change in the legal industry and educate in-house counsel on how they should adapt to it.

LLMs could never become a substitute for the skills external panels bring when handling complex deals (such as mergers and acquisitions) as they have to manage a number of current and emerging risks, Dickason pointed out.

“But LLMs could easily handle tasks in lower-risk environments where the questions are relatively simple to answer and the answers can be fact-checked quite quickly,” he said.

For example, the generative AI solution Lexis+ AI uses retrieval augmented generation (RAG) to increase the accuracy and authoritativeness of LLM output by cross-referencing a trusted source of knowledge. The tool refers to specific, high-quality legal content sets to ground its responses in truth and reduce the risk of AI hallucination (where a model generates false information).

Dickason said that while the product was released in the US over six months ago, it has only recently been released in Australia.

In the US, the legal profession has been using the product for drafting, summarising, and comparing large numbers of legal documents, while Australia “is on the cusp of this”, he added.

“I think in-house lawyers are increasingly going to undertake these tasks themselves and won’t need to go to external panels,” Dickason concluded.

To hear more from Greg Dickason about what generative AI is and how in-house counsel could harness its power to propel their professional journey, come along to the Corporate Counsel Summit.

It will be held on Thursday, 2 May 2024 at The Star, Sydney.

Click here to book your tickets and don’t miss out!

For more information, including speakers and agenda, click here.