Skip to content

Our Privacy Statement & Cookie Policy

All Thomson Reuters websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.

Artificial Intelligence

Professional-grade GenAI: Four common questions, answered

· 6 minute read

· 6 minute read

Legal professionals have heard a great deal about how AI can benefit their practice. But what should these professionals know before they incorporate GenAI into their work?

← Blog home

Jump to ↓

1. What is professional-grade GenAI?

2. What developments have made professional-grade GenAI possible?

3. How can professional-grade GenAI be used?

4. How can professionals evaluate professional-grade GenAI?

Moving forward with GenAI

 

The Wall Street Journal recently met with Jake Heller, Head of Product, CoCounsel at Thomson Reuters, about professional-grade GenAI. It’s a subject that Heller knows intimately. Before joining Thomson Reuters, he was CEO of Casetext, a developer of AI-based technology for legal professionals. When Thomson Reuters acquired Casetext in mid-2023, Heller came on board to lead CoCounsel’s ongoing development. CoCounsel is an GenAI-powered legal assistant that delivers document review, legal research memos, deposition preparation, and contract analysis, among other capabilities.  

While many legal professionals have already begun using GenAI, many are still wary. They worry, for instance, about this young technology delivering unreliable information and exposing sensitive data. But as Heller argued in the Wall Street Journal forum, professional-grade AI solutions can provide reassurance along with its benefits. With professional-grade AI attracting increasing interest, here are answers to four common questions that legal professionals may be wondering about. 

 

1. What is professional-grade GenAI?

Professional-grade GenAI refers to advanced artificial intelligence capable of delivering high-accuracy results, making it suitable for high-stakes professional environments, including legal work. 

Professional-grade GenAI must be:  

  • Developed by technology teams with deep AI expertise.  
  • Continually and consistently tested and authenticated by experts in the user’s field. In the case of CoCounsel, this expertise is rooted in Thomson Reuters’ century-plus history of creating reliable legal content, such as its Westlaw online legal research database. 
  • Secure and private, interacting with external information sources so that the user’s data is protected from hackers and other outsiders. 

In the conversation with Wall Street Journal, Heller noted that we should distinguish professional-grade GenAI with current consumer applications of the technology. People are using ChatGPT and other AI tools for content creation, photo editing, design, and gaming. But as Heller noted, professionals need more rigorous tools. 

To be truly professional-grade, a GenAI solution “needs to be tested [by the developer] so that it doesn’t hallucinate,” Heller said. Hallucinations are incorrect or misleading results generated by AI models. They occur when AI’s large language models (LLMs), which process, understand, and generate human language, produce false information in response to user prompts. Professional-grade GenAI is more reliable because it “draws from real sources like the law itself,” Heller said.  

 

2. What developments have made professional-grade GenAI possible?

From Heller’s perspective, the breakthrough that more than any other factor made professional-grade AI viable was the arrival of GPT-4 in March 2023. This new version of the GPT LLM suddenly allowed Casetext—“in a weekend,” as Heller put it—to make real what had seemed up to that time impossible. GPT-4 made development of capabilities such as research and document review much quicker to achieve.  

During its product development process, Casetext tested versions of GPT by having them “take” a standardized bar exam. GPT-3.5 scored in the 10th percentile. GPT-4 scored in the 90th percentile. This demonstrated to Heller and his team that “this is something that is reading and understanding and writing at a postgraduate level.”  

What made this big leap forward possible? “The simplest way to put it is, these AIs aren’t just generative,” Heller said. In other words, these technologies do more than simply generate and deliver requested results from existing data.  

“They’re also thinking machines,” Heller added.  “And that unlocks a lot of capability.

 

3. How can professional-grade GenAI be used?

Heller noted that GenAI’s current iteration “unlock[s] so many capabilities, whether it be helping you do legal research…reading through cases and statutes and regulations, pulling out key insights,” among many others. CoCounsel, for instance, can draw upon the reliable legal data in Westlaw and quickly deliver insights across documents, cases, and rulings.  

“It’s almost infinite the number of things you can do now that you have GenAI that can read and write and understand at this level,” Heller said. “It’s working at the speed of technology.”

  

4. How can professionals evaluate professional-grade GenAI?

For legal professionals who are intrigued by these possibilities, conducting thorough due diligence is, of course, essential. Heller noted that assessing a potential GenAI-powered tool is as complex as it is necessary. “If I were to give one piece of advice,” he said, “it would be to set up a bunch of rigorous tests.” For instance, legal practitioners could have the GenAI tool read and review thousands of documents and determine the accuracy of the analysis it delivers.  

During the Wall Street Journal discussion, an audience member asked Heller whether GenAI truly has the capability to interpret documents and summarize their intentions. Heller replied that it can–if the professional using it is able to train it correctly. From there, the technology can “learn” and continue to improve. “I wouldn’t undercount these machines…in terms of how deep their intelligence really is,” Heller concluded.  

 

Moving forward with GenAI

Given how rapidly GenAI is developing, professionals will want to conduct due diligence on the vendors as well as on the solutions they offer. In addition to meeting the reliability requirements Heller discussed, the best professional-grade solutions need to:  

  • Intelligently and seamlessly handle workflows, not simply perform a series of specifically requested tasks 
  • Provide partnership, not just product, with ongoing support resources. 

For AI to provide maximum value to legal professionals (and other types of professionals), solution providers should demonstrate that they can keep pace with new LLMs, optimizing existing GenAI-powered products and developing new ones. In addition, the rapid pace of AI development has made it challenging, but also more important than ever, for professionals to stay informed about regulatory and ethical issues regarding AI use to maximize its benefits and minimize the risks.  

Professional-grade AI is quickly establishing itself as the standard for AI solutions for legal professionals, as well as practitioners in industries including tax and accounting. Learn more about how professional-grade GenAI is evolving and how it is positively transforming the way legal professionals work by accessing the complete Wall Street Journal discussion with Jake Heller 

CoCounsel

CoCounsel

Bringing together generative AI, trusted content and expert insights

Meet your legal AI assistant ↗

 

More answers