WHITE PAPER

Generative AI for legal professionals: What to know and what to do right now

The revolution in generative AI (GenAI) is quickly transforming business as we know it, and legal is helping lead the charge. Within the legal industry, AI now touches every role, both front- and back-office. New generative AI-based legal tools can support multiple roles within a law firm or legal context, including marketing, administrative, and support departments. While firms will implement such tools on different timelines for different uses — some extremely rapidly and some more slowly — there is no question that GenAI is here to stay as a feature of legal practice.

For the traditionally risk-averse legal profession, this can feel a bit uncomfortable. However, the 2023 Thomson Reuters Future of Professionals Report found that legal professionals are beginning to recognize how generative AI can help their practices improve operations and increase efficiency, which are priorities for the profession. “For law firms,” the report states, “professionals see the opportunity for AI to enable productivity in two ways: saving time by conducting large-scale data analysis and performing nonbillable administrative work with greater accuracy. Both productivity enhancements highlight the fact that these tasks can be done faster and better, allowing fee earners to make greater billings.” 

The key to using GenAI to facilitate legal work is an industry-specific solution designed by legal insiders and trained on the highest-caliber legal content. Such best-in-class, industry-specific tools can substantially support law teams’ daily workflows while significantly reducing the inherent risk of using this next-generation technology.

The current state of generative AI in the legal landscape

The rise of generative AI is upending aspects of legal work in a way that previous AI tools have not. AI models have been used for years to streamline workflows and back-end processes — such as improving search functions and data mining — but new generative AI tools offer substantively enhanced capabilities.

Legal AI tools entering the market today aren’t just for general use; they are industry-specific and trained using vetted, specialized, and proprietary legal knowledge. These tools can rapidly respond to plain-text queries with extremely relevant outputs, including research, synthesis, and analysis. Additionally, their ability to generate plausible text can assist lawyers in delivering actual legal work product, including core output like contracts and legal summaries. This ability will become increasingly useful as the capabilities of large-language models improve.

Attitudes toward this evolution within the profession are mixed. While 82% of legal professionals believe they can apply generative AI to legal work, only 51% think it should be, according to the Thomson Reuters Institute’s ChatGPT and generative AI within law firms report. In practice, according to the report, only 5% of legal professionals state that their firms have plans to use generative AI, while 60% say their firms currently have no such plans. About one-third of firms are reportedly on the fence about whether to use generative AI.

This hesitation is understandable. The top two concerns voiced by legal professionals in the ChatGPT report are that GenAI will take the place of professional legal judgment and replace lawyers and that generative AI outputs are too error-prone and opaque to be helpful in practice.

There is indeed good reason for lawyers to be skeptical of large-language models. This was illustrated most prominently by the May 2023 instance of lawyers citing fabricated cases in a court filing for a personal injury suit, a result of “hallucinations” produced by ChatGPT that were included in the filing without proper vetting. With such headlines shining a spotlight on the technology’s shortcomings, it is little wonder the ChatGPT report found that 72% percent of legal professionals believe that generative AI should be limited to non-legal work within their firms.

However, the belief that generative AI is bound to be more problematic than helpful for the full range of legal work is based on an incomplete understanding of how the legal profession can best use AI. Since legal generative AI is intended to augment the work of legal professionals, it can vastly improve lawyers’ efficiency, speed, and creativity if used in the right ways. The quality of legal AI tools is vital — as is the practice of viewing such tools as assistants whose work always needs validation.

GenAI still requires the human element

As the ChatGPT hallucination story demonstrates, generative AI is not intended to be used to create presentation-ready legal outputs. Legal professionals should approach these tools as they approach the work of a skilled legal assistant. Although AI can make the human-centered work of law more streamlined, lawyers must apply their expert knowledge by fact-checking AI-produced assertions, building on GenAI research suggestions, rewriting its stilted content, and enhancing the legal arguments it constructs. Like professionals in other fields using these next-generation tools, legal professionals should think of generative AI as a research and ideating aide that works at the speed of light but doesn’t possess even a fraction of a lawyer’s knowledge.

Some positive misconceptions about generative AI are also common in the legal profession — and these can also impede the effective use of generative AI in legal work. For example, some professionals believe that training a large-language model is simple and fast — and these tools can deliver helpful results with minimal training. The truth is it takes extensive work to train large-language models to be beneficial to legal professionals, requiring data scientists and large volumes of legal materials that provide context-specific language and concepts.

In considering the use of generative AI for legal work, it’s essential for law practices to note that there is a vast difference between free, public-access generative AI tools like ChatGPT and industry-specific, highly refined products like the Thomson Reuters AI-Assisted Research on Westlaw Precision. Third-party providers like Thomson Reuters are taking on the work of training such tools on lawyers’ behalf, putting their data scientists and legal experts to work on their extensive legal large language models.

Why legal professionals should embrace generative AI

Generative AI tools for the legal profession are specifically designed to help legal teams perform dozens of tasks more efficiently in their daily work.

Below are some of the benefits for the legal professional.

Efficiency and speed. Generative AI uses the power of automation to help legal teams perform a range of tasks more efficiently and quickly. Examples of work that generative AI can help with include summarizing legal content, planning for depositions, synthesizing and reviewing content, and conducting contract analysis. Such tasks take a significant amount of time when done manually, so the extreme speed of GenAI leaves lawyers more time to thoughtfully assess and build on what the large-language model produces. More efficient legal research means minimized nonbillable time and reduced administrative burden.

Legal-specific support. Built-for-purpose generative AI is significantly different than GenAI for wider use cases. It is trained, built, and optimized for legal use and is held to higher standards around data, privacy, training models, and content. For example, some generative AI-enabled tools contribute to legal research and support various applications that seamlessly integrate into law firms’ workflows. Such integration allows teams to complete rote tasks specific to legal work far more quickly and with a high degree of confidence that the outputs will be helpful.

Satisfied clients. Legal professionals face client pressure to adopt new technology and deliver more value to automate rote back-office functions. Clients expect their lawyers to be on the innovative side of technology and to use the best tools available. Generative AI helps legal professionals deliver on those client expectations faster and more efficiently from a better starting point, thereby reducing friction, cost, and other impediments to providing optimal customer service. As a result, legal professionals who use generative AI-enabled tools strategically can pass on the benefits of the resulting time and cost savings to clients. 

Reduced outsourcing. In-house counsel can use generative AI-enabled tools to perform critical tasks faster, reducing the work they need to outsource. Doing less outsourcing can substantially reduce costs and remove bottlenecks that slow progress throughout an organization or government agency. Reducing costs helps in-house law departments move toward the goal of working as a revenue driver instead of a cost center for their organizations.

Revenue generation. In-house counsel in the private sector can use this technology to help generate revenue on behalf of the organization, thereby allowing their department to shift away from being a cost center. One example is speeding up the contract analysis process to help the business close deals faster. Additionally, when highly skilled legal professionals are freed from rote tasks, they can spend more time on high-value activities that drive the business forward and make the best use of their training and knowledge. 

What drives your large-language model?

Lawyers have been reluctant to adopt GenAI tools due to a lack of trust and fear of the increased risk that these tools might introduce. A majority of legal professionals (62%) say their firm has concerns about the dangers of using generative AI in the workplace, and 80% of partners and managing partners believe there is concern around risk, according to the Thomson Reuters Institute’s report mentioned previously. Only 2% say their firm has no risk-related misgivings about the technology.

However, much of the perception of generative AI’s risk is based on misconceptions and a lack of knowledge of the built-for-purpose tools available to the legal profession. While public-facing GenAI tools like ChatGPT use open-source, stale data with few security and privacy controls, legal-specific generative AI tools use high-quality data in a safe and secure environment. Professional tools like those from Thomson Reuters use closed, secure data sets in an environment that maintains privacy, and models are trained on trusted content that is not freely available to the general public.

Why the right content matters

The only way for legal professionals to trust the output of the AI tools they use is for those large language models to be powered with trusted content specific to their professions and with robust security and privacy controls built in.

Law firms that want to leverage the massive advantage in efficiency and speed that AI provides need a best-in-class tool built on trusted data — and they need to mandate its exclusive use within their firms. Yet most legal teams today are doing little to counter the risk of employees using whatever generative AI tool they prefer. According to the Thomson Reuters ChatGPT report, only 15% of survey respondents said their firms had advised employees to avoid unauthorized use of generative AI tools, and only 6% reported that their firms had banned unauthorized use.

How legal-focused GenAI tools help law firms and legal departments

Generative AI tools custom-built for law practice provide targeted benefits for legal organizations beyond what generic GenAI solutions can offer. There are numerous benefits, many of which contribute directly to the firm’s or department’s ability to serve its clients more efficiently.

More efficient legal research

Generative AI provides a jumpstart on legal research by reducing the time legal professionals must spend sifting through and summarizing content. These tools can produce, in moments, an informative version of research that would have previously taken hours or days. Lawyers can then apply their expertise to refining the results to ensure the research output is high quality, thorough, and accurate. This jumpstart is like beginning a journey at its midway point, leaving the legal firm and its client far more time and energy to reach the end successfully.

Faster synthetization and analysis

Legal professionals can use generative AI to produce a range of useful outputs, including content summaries, meeting agendas and transcripts, contract analysis, and document review. The synthesis and analysis required to create these products happen in the blink of an eye, leaving users time to check, enhance, refine, and rewrite where necessary. The range of assistance offered by legal-focused generative AI improves the legal firm’s or department’s output by facilitating the legal professional’s ability to apply their higher-level expertise to the work at hand.

Improved knowledge management

Generative AI can help legal teams stay organized and share information via cross-functional tools. Better and faster methods of saving, indexing, identifying, and disseminating lawyers’ prior work and collective expertise help legal firms and departments solve legal and business problems more effectively. Legal teams can use these tools to rapidly draw upon the store of knowledge, allowing them to enhance their analysis without extensive or frustrating searching. The work that goes into knowledge management is typically not billable, so making it faster and easier helps the legal firm or department save costs and free up professionals’ time for billable work.

Enhanced onboarding and learning

Legal organizations benefit from efficient onboarding and ongoing skills development, both of which GenAI can enhance. Legal-focused generative AI tools help users get up to speed quickly without extensive training, whether they are new to a firm or gaining knowledge about an unfamiliar area of law. These tools also help users learn new skills hands-on as they work, reducing the need for as many formalized skill-development interventions, saving time and money.

Plain-language prompting to navigate complexity

Legal matters are often complex, making research, knowledge management, synthesis, and analysis difficult. Those unfamiliar with a given area of law may struggle to know where to start or which language is most applicable in searching for resources. GenAI tools that can be prompted using plain-language queries allow users to access faster answers to complicated legal questions and rapidly organize information and precedent to help them develop successful arguments.

Trustworthy security and privacy

Data privacy and security are paramount concerns in the legal industry. A professional tool licensed to a particular institution or firm provides a far more secure work ecosystem than a public-facing tool with few or no data controls. While legal professionals must always use caution regarding which data they use with AI tools, they can input a range of proprietary data with far less risk in a professional, licensed tool. Additionally, since legal-specific generative AI tools are trained on high-quality legal content, legal organizations can be confident that their output will be more trustworthy and accurate than other large-language models.

Finding the right AI partner

Thomson Reuters has been a trusted resource within the legal industry for 150 years and is now a leader in legal AI. Almost 4,500 law firms, corporate departments, and government agencies use tools from Thomson Reuters, and our labs have been developing solutions driven by large-language models since the early days of their development.

We work hand in hand with partners and customers to build solutions iteratively, a method that ensures those solutions most exactly meet their needs.

We train our AI tools on best-in-class content created by law experts, which makes them far more useful and trustworthy to legal professionals than other GenAI tools on the market or those available to the public for free. This intensive training helps mitigate problems like hallucinations and missing content in responses and enhances users’ trust in and experience with the tool.

Powerful tools for legal minds

While generative AI is a powerful tool in any form, customizing its functionality to a particular context amplifies its usefulness. Legal audiences need legal-specific generative AI to fully benefit from this technology, which can help speed up research and analysis, better satisfy customers, reduce costs, and boost revenue.

Made-for-purpose GenAI tools can add tremendous value to legal work, but they won’t replace legal professionals. Generative AI technology can augment the efforts of legal professionals by increasing efficiency, jumpstarting research, facilitating synthesis, and augmenting analysis — but humans’ legal expertise and judgment will remain the core driver of the profession.

Generative AI is a powerful tool — but as with any tool, it’s most beneficial when wielded by a skilled professional. As this technology develops, forward-looking legal teams will benefit from learning about its risks and benefits from an expert with a century and a half of legal industry experience.

Learn more about AI @ Thomson Reuters, and sign up for industry-leading insights and updates on this fast-developing topic.

Get trusted answers on AI

Sign up to receive industry-leading insights, updates, and all things AI @ Thomson Reuters — delivered monthly to your inbox