Why The Future of Legal AI is Trusted Legal Work

LexisNexis® CTO Greg Dickason joined Lawyers Weekly's LawTech Talks podcast to discuss why the future of legal AI lies in integrated, governed systems and what that means for New Zealand legal professionals.

Most conversations about legal AI focus on what it can do.

In a recent Lawyers Weekly LawTech Talks interview, LexisNexis Chief Technology Officer Greg Dickason, however, believes the more important question is ‘what it can produce’.

"Whereas the AI of the past would give me the answer," Greg told Lawyers Weekly host Jerome Doraisamy, "the AI of the future helps me build the output; the final thing, which is my client deliverable. You want it trusted, you want it authoritative, you want to be able to verify the links, but you also want it to give you really high-quality outputs."

This distinction matters.

Legal work is not simply about generating text. It involves advice, advocacy, risk assessment, and professional judgment. As AI becomes more capable, the profession is increasingly focused on a different question: can lawyers trust, verify, and defend the work it produces?

Why capability alone isn’t enough

For Greg, the challenge is not whether AI can help lawyers work faster. It is whether the outputs can be relied upon.

"AI is very good, but it's not perfect. The modern models have been tuned around trying to make you feel good - they give you good answers, they try to give complete answers. And in some cases, that means they make up facts. In a lot of cases, they make up facts."

For New Zealand practitioners, this is not a theoretical concern. Around the world, courts are increasingly confronting examples of hallucinated citations, fabricated authorities, and AI-generated content that cannot be verified.

As adoption accelerates, many lawyers are moving beyond experimentation and asking a more practical question: how can AI be used in a way that supports professional obligations and reduces risk?

Greg sees this as the next phase of legal AI adoption.

"Increasingly, lawyers are saying, I need to make sure this is grounded, this is authoritative. And the reason why that's happening is because that they're increasingly seeing penalties... because in some cases, unfortunately, lawyers have put work that is hallucinated in front of the court. We're now in that phase where lawyers are saying, hang on, I need to trust this content."

The rise of integrated legal intelligence

At the same time, legal professionals are facing another challenge: fragmentation.

Research happens in one system. Drafting happens in another. General AI tools sit outside existing legal workflows. Documents, emails, and matter information are spread across multiple environments with different security and governance controls.

Greg believes the future of legal AI is not another standalone tool. It is an integrated legal intelligence environment.

When Jerome asked Greg to define what integrated legal intelligence actually means in practice, his answer was immediate:

"As a lawyer, you can go from thinking, I'm going to go over here to ask a question, and then back to over here to do my work, to actually my work and the AI and the ability to ask the questions all integrated. So I can get a message from my client — an email — dump that into the AI, do some research with the AI, and then create almost close to a final output. Ultimately, I'm getting a Word document, a PowerPoint, maybe a complex set of those that really meet my client's needs. That's what I mean by integrated legal intelligence. It's integrated throughout your workflow. It's integrated throughout your day. It's not something you go to and leave."

"It's integrated throughout your workflow. It's integrated throughout your day."

From legal research to completed legal work 

That vision is beginning to take shape through Protégé™ Work, arriving in New Zealand in early July.

Rather than generating a single response, Protégé Work helps legal professionals move from a legal objective to a completed piece of legal work. It creates a structured plan, allows the lawyer to review and refine the approach, and then executes the work through a series of connected legal tasks.

As Greg explains:

"There's a new product called Work that's coming out... and that product is effectively very similar to the way Claude's Co-Work works, from Anthropic, but it's grounded in our content. It can pull all the skills... it gives you a lot of already predefined workflows that you can use, ones that you can import and build yourself. And it produces that final output - the Word document, the PowerPoint, etc. So, it's the workflow, the final work output, but grounded in authoritative content."

The mechanics are handled. The judgment stays with the lawyer.

Research, drafting, analysis, and review come together within a private, secure environment grounded in authoritative LexisNexis content and organisational knowledge.

What this means for New Zealand practitioners

Greg's advice for legal professionals considering AI is straightforward:

"The big question is, how am I able to get more productive, be able to do higher value work by using AI? And then where is that AI coming from? Is it grounded? Is it authoritative? Can I trust it? And most importantly, can I use it to verify - can I see and verify the sources that it quotes?"

These questions are becoming increasingly important as AI moves from experimentation into everyday legal practice.

The future of legal AI will not be defined solely by capability. It will be defined by trust, verification, governance, and the ability to produce legal work that professionals can confidently stand behind.

Protégé has been designed with those same principles in mind.

Listen to the full conversation with Greg Dickason on Lawyers Weekly LawTech Talks here.