[ad_1]
Ed. note: This article first appeared in the Winter 2023 edition of ILTA’s Peer to Peer magazine. For more, visit our ILTA on ATL channel here.
In 2022, no one could have predicted that Generative AI (Gen AI) would be the most talked-about technology of 2023, yet only a few months into the new year, ChatGPT took the world by storm. It quickly became mainstream with its dazzling ability to answer questions, create and summarize content, and accelerate mundane tasks.
The legal industry was no different. Although Large Language Models (LLMs) and extractive AI had been around and in use for years, and Gen AI models were already being evaluated for industry use, ChatGPT skyrocketed our excitement for new Gen AI-powered technologies and potential use cases that would improve the practice of law and make our lives easier. Vendors and law firms began touting their new Gen AI technologies. But all too soon, emerging stories about potential copyright violations, egregious hallucinations, and blatant biases dampened our enthusiasm and made us question its benefits.
At Legalweek in March 2023, a three-hour, standing-room-only Gen AI workshop showcased the industry’s strong desire for more insights, information, and answers. Since then, various incidents have shown us the potential ramifications and direct consequences of using an unproven AI technology that was not “professional grade.” Under normal circumstances, such a negative outcome would permanently disable any other legal technology, not Gen AI. Instead, it galvanized vital stakeholders to develop responsible development and use principles, guardrails, and guidelines that protect attorneys and clients, as well as new strategies to create trustworthy AI experiences. By the end of 2023, Gen AI’s momentum and outlook were as strong as ever.
So, where is Gen AI headed in 2024 and beyond? History has shown that lawyers want friction free, easy-to-use, streamlined, and interconnected applications. We believe that 2024 will herald the rise of secure, domain-specific, integrated Gen AI solutions that deliver unprecedented increases in efficiency, productivity, and knowledge, and empower attorneys and clients to achieve their business and legal objectives.
Current Gen AI Challenges
Taking a step back, it’s essential to understand that while developments are accelerating to meet the demand, delivering secure, domain-specific, integrated Gen AI solutions will not happen overnight. Many obstacles must be conquered — on both the model development and technology delivery side — to achieve this goal. Separately, there’s the question of which models to use for which use cases — as no one model currently under development exceeds performance expectations for all tasks. And then there’s rebuilding trust among wary attorneys, clients, and judges.
As we’ve seen, LLMs are not “professional grade” enough for complex, knowledge-worker applications right out of the box. Specifically, the legal industry should not use any model trained by opaque, open-web data full of inaccuracies and biases. By themselves, LLMs cannot validate the accuracy of their output; they can merely infer what they believe to be accurate based on the unvetted information they ingest. This leads to plausible-sounding Gen AI “hallucinations,” incorrect or misleading information, and general mistrust of the technology, so much so that judges must ban its use or compel lawyers to attest to the integrity of the citations. Corporate attorneys are increasingly requesting greater transparency and the ability to determine if, when, where, and how AI is used in tools they use to assist in creating their legal work product.
Another challenge is setting realistic expectations for what Gen AI can and cannot do (or should not do) and delivering upon them. New technologies that fail to meet the requisite efficiency, productivity, knowledge, or process improvements (beyond what’s currently available) or exceed attorneys’ exceptionally high standards for privacy, security, and compliance are quickly discarded. According to Gartner’s 2023 Hype Cycle for Emerging Technologies, Gen AI has already reached the Peak of Inflated Expectations and is headed towards the Trough of Disillusionment — a period of waning interest as initial experiments and implementations fail to deliver. To reach the Slope of Enlightenment, when a technology’s benefits are more clearly understood and materialize, solution providers must rapidly innovate and address the many new challenges that Gen AI creates. We believe growing customer interest and demand and increasing vendor competition will help Gen AI accelerate through this phase. Indeed, Gartner predicts Gen AI will reach transformational benefit (i.e., the Plateau of Productivity) within two to five years — much faster than the traditional seven to ten years for most new technologies.
Building Trust with Domain-Specific Responses
To overcome these challenges and build trust among users, solution providers need to clearly demonstrate the accuracy and transparency of their Gen AI applications and services, backed by verifiable authority, and minimize instances of invented content.
Achieving this requires the creation of legal domain-specific LLMs that prioritize the improvement of model output. This is done using subject matter experts — attorneys — to fine-tune models for specific legal use cases; prompt engineering that analyzes a customer’s question and adds additional instructions to the model; and integrating vast amounts of caselaw, legal data, news, and other content capabilities using Retrieval-Augmented Generation (RAG) to augment models. RAG extends a model’s capacity by connecting it to an external knowledge source. Organizations with access to this high-quality content and pristine data are better positioned to jointly partner with LLM creators to develop models for legal industry use. Our first-hand experience using language models dates to 2018 with Google BERT. Today, we’re working directly with LLM creators and trusted cloud providers to develop faster, more accurate, transparent, and secure Gen AI offerings, giving users peace of mind when leveraging these advanced technologies.
Ensuring Privacy, Security, and Compliance Globally
Similarly, Gen AI privacy and security issues remain a high priority for legal professionals. Of concern is that sensitive inputs or uploaded documents might be searchable or used to train core LLM models. To alleviate concerns, solution providers are working with cloud hosts like AWS Bedrock and Microsoft Azure to create “walled gardens” — secure cloud environments that ensure confidential data stays within the system and is not used to train the model. Data privacy best practices dictate that all user data should be encrypted in transit for added security. Other privacy safeguards should be built into Gen AI applications, such as purging uploaded documents and prompt histories — manually or automatically at the end of each session or a defined period of inactivity.
Regulatory compliance is equally concerning, especially among multinational organizations looking to implement Gen AI solutions globally. Since global jurisdictions may have different privacy and data-sharing rules and expectations (e.g., the EU’s GDPR, etc.), organizations must ensure that the Gen AI solutions they implement comply with all applicable laws before rolling them out to their global employees. Solution providers that already have a global footprint will have an advantage over smaller providers in helping global organizations comply with local regulations. This compliance could be leveraged by organizations to market globally compliant data solutions to prospective customers and win new business.
Gen AI will facilitate the inspection and analysis of customers’ proprietary data — but this needs to be done safely and securely.
The Case for Integrated Gen AI Applications
Demand for legal Gen AI applications is already unprecedented. According to MarketResearch.Biz, legal industry spending on Gen AI was projected to reach $65 million in 2023 and surpass $675 million by 2032, with a CAGR of 30.7% from 2023 to 2032. However, as we’ve seen in the past with other transformative technologies, the initial rush to purchase point solutions generally leads to increased friction, frustration, and dissatisfaction upon discovering their lack of interoperability with existing applications, data sources, workflows, and business functions or their inability to support global legal operations.
Since a significant portion of a typical attorney’s day is spent in the Microsoft environment (e.g., Word, Outlook, Teams, etc.), organizations will look for Gen AI solutions that seamlessly integrate with these applications and elevate their functions. For example, Gen AI-powered conversational search, summarization, and document drafting and analysis pair well with Microsoft Word when creating briefs, motions, contracts, and other legal documents. Similarly, natural language chatbots that can answer basic legal queries, engage in multi-turn, iterative conversations, and automate work intake requests via Microsoft Outlook or Teams can improve legal team productivity.
Integrating with existing data sources — including proprietary client or corporate data and third party sources — will be critical to maximizing the potential and capabilities of an organization’s Gen AI investment and minimizing instances of invented content. Most point solutions cannot access data from competitive third-party providers, which limits their effectiveness.
Similarly, large global organizations will benefit from the synergies and efficiencies that deploying one integrated Gen AI platform creates. The ability to seamlessly access proprietary corporate or firm-wide data, translate it into the local language, enhance it with localized information (e.g., court and financial records, news, etc.), and leverage it to draft new documents — all under the governance of local compliance regulations and data security is something that multiple disparate point solutions cannot quickly provide. Additionally, if attorneys travel frequently between global offices, working comfortably in a consistent Gen AI environment will improve efficiency and productivity.
Cost and Other Considerations
While implementing individual point solutions might seem to be more cost-effective in the short-term than purchasing a fully-featured, integrated Gen AI solution, many metrics beyond cost should be considered, including user efficiency, productivity, and satisfaction, training requirements for multiple point solutions, long-term technology roadmaps and other IT considerations — all of which can impact the total overall cost of implementing and adopting Gen AI. For example:
Model Upgrades: As models improve and change or new ones are created, organizations might find individual point solutions unable to keep up with the latest advances. An integrated Gen AI platform framework would allow easy model upgrades or replacements as technologies improve, giving organizations peak performance with little disruption.
Continuous innovation: As fierce competition drives new technological and product innovations, first-generation point solutions implemented today could be leap-frogged by other solutions tomorrow, necessitating a disruptive switch in solutions and/or providers to gain the desired features and performance. As use cases evolve, a comprehensive Gen AI platform will be better able to implement new features without disrupting established workflows or existing applications and data integrations.
M&A: As we’ve seen with other legal start-ups, innovative Gen AI startups will likely be acquired by more prominent players to maintain competitiveness, gain market share, and develop new tech capabilities and talent. Gen AI’s rapid descent into Gartner’s Trough of Disillusionment could hasten this process, especially for smaller solution providers that cannot keep up with innovation or are unprepared to weather the inevitable shakeout. An organization’s investment in Gen AI point solutions, including building workflows and processes around them and training employees to use them, might one day be disrupted if one or more point solutions are acquired.
The role of the attorney: As Gen AI advances over time, the role of the attorney will naturally evolve. Gen AI will transform legal work, not replace it, and attorneys will benefit from productivity gains in essential ways. Ultimately, Gen AI will enable attorneys to spend more of their time on higher-value work.
Undoubtedly, the efficiency and productivity benefits of Gen AI have captivated the imaginations of the legal industry. The industry has shown a great and rapid willingness to embrace Gen AI. A recent international survey showed that nearly all attorneys (92%) believe Gen AI will impact the legal profession, with 47% believing it will have a transformative effect, and 43% either currently using or plan to use generative AI in their legal work.
As we continue to push the boundaries of Gen AI and LLM capabilities, the industry will experience a flood of new solutions that address many of the key pain points experienced by practicing attorneys — but also create new challenges for their organizations that must be addressed. As always, cost will be a crucial factor that determines whether or when organizations adopt Gen AI and which solutions they choose. However, forward-looking decision-makers who remember past technology hype-cycles will look beyond cost and factor in broader criteria such as seamless application and data integrations; enhanced efficiency and productivity; increased privacy, security, and compliance; improved workflow and globalization capabilities; easy future upgrades and roadmap planning; and more. All these criteria converge toward a single, secure, interconnected Gen AI platform that can support multiple use cases and evolve to meet our industry’s changing needs and rapid technological innovation.
Jamie Buckley is Chief Product Officer for LexisNexis Legal & Professional. In this role, Jamie is responsible for the software products built for our amazing customers. He focuses on delivering successful and innovative products and platforms that serve customer needs around the globe.
Jeff Reihl is the Executive Vice President and Chief Technology Officer (CTO) for the global legal business of LexisNexis. In this capacity, he is responsible for global technology strategy, bringing together the company’s applications, product platforms, and business systems to deliver LexisNexis world-class content to its customers in innovative ways.
[ad_2]
