The Generative AI Revolution Is Here
By: Daniel J. Cravens
The Generative AI revolution is at the gates of your law firm. You are skeptical for good reason, but your associates are already adopting available tools in their workflows. How can you best reap the benefits while minimizing exposure?
Attorneys, especially partners with a little gray in their hair, can be notoriously slow to adopt new technologies. I recall partners at my first law firm in the late 1990s urging me to learn dictation instead of using a word processor long after computers had made that workflow obsolete. One storied litigation partner at the firm permitted associates to use computerized research services in their work for him, but only if they double-checked their conclusions using books from the firm’s library, which incidentally occupied an entire floor of expensive San Francisco real estate. But attorneys are skeptical of new technology for good reason. We are legally and ethically required to provide our clients with competent legal advice. The consequences to early adopters of misunderstanding this new technology can be severe. Take the case of Steven Schwartz, a New York attorney with 30 years of experience, who may end up facing sanctions for submitting an AI-generated brief replete with errors. Mr. Schwartz could be applauded for his recognition of the transformative potential of Generative AI, even as he reminds us that the attorney who signs the pleading is ultimately responsible for its contents.
Generative AI Is Already Being Used In the Industry and Probably in Your Firm
According to Bloomberg Law’s State of the Practice Survey 2023, half of all attorneys see potential in using AI to draft, review, and summarize legal documents. A solid 75% view it as appropriate for legal research. The implications? AI isn’t just coming—it’s already embedded in legal practices to one degree or another.
Further, the Above the Law and Wolters Kluwer Survey from April 2023 provides more food for thought. Nearly two-thirds of respondents are betting on Generative AI being the game-changer in the next half-decade. Additionally, 80% are convinced that it’s set to overhaul legal research and routine tasks.
Generative AI tools are already available to attorneys for free or at low cost. GPT-3.5 is a free, but notoriously limited, tool already available to attorneys. The far more powerful GPT-4 is also available to anyone for a nominal charge.
Both Thomson Reuters and LexisNexis are launching Generative AI tools this year. These tools promise to address clients’ concerns about maintaining confidentiality of client data and competency of results. These sanctioned tools may be safer and more efficient for most use cases in the legal field. But they are sure to be expensive, creating an opportunity for deep-pocket firms to outperform less well-funded competitors, if they are adaptive enough to take advantage of the technology.
A recent white paper authored by John Villasenor of the Brookings Institute suggests that law firms may be able to use Generative AI to increase efficiency for a number of important tasks. For example, AI applications have the potential to transform traditional document review and production, which is a vital function hated both by the associates who spend countless vision-blurring hours reviewing documents and by the clients who pay eye-popping fees for the service. An AI tool that performs this task faster, more cheaply, and potentially at a higher quality is a win for everyone. Although not yet mainstream, there are potential applications for Generative AI tools that will transform legal research and document drafting.
Count On Generative AI To Disrupt The Legal Industry Over The Next Five Years
Bill Gates recently wrote that Generative AI is disruptive across industries, education, health, and government. Gates predicts that there will be “an explosion” of new commercial applications for Generative AI. Disruptive applications in health and education have received much attention in the news. Although less well publicized, disruptive innovation in the legal field has already begun.
There are several start-ups producing products that promise to help with drafting routine documents. AI startup Harvey is marketed as Legal AI which will collaborate with elite law firms to transform every legal task from research to drafting. Yet, Harvey is careful to emphasize that its product is intended to amplify the productivity of all legal “knowledge workers”, a population that pointedly includes more than barred-attorneys. Rally Legal, another startup, is testing a product called “Spellbook” that helps draft contracts more efficiently. It remains to be seen whether these products are better marketed to attorneys as efficiency-enhancing tools or to clients as an alternative to traditional legal services. The latter risk is highlighted by a Goldman Sachs report stating that 44% of its legal tasks could be tackled by knowledge workers utilizing AI products. Even where these AI products are currently marketed as tools for attorneys, it seems likely they will soon evolve into competitors as the technology and legal spaces mature. Outside competitors have an advantage over law firms because they can raise money in the capital markets.
Some firms will lose business to these new tools, but there will be offsetting opportunities as well. For example, there is an opportunity in the emerging field of AI law to advise start-ups designing AI tools and to represent clients who find themselves in legal peril based on their use of these tools. Moreover, the availability of Generative AI legal tools is likely to increase litigation by pro se plaintiffs. Robolawyer, for example, offers an AI product targeting consumers called Robo Revenge AI that will prepare lawsuits against illegal robo-callers who violate the Telephone Consumer Protection Act. Another startup, Reveal AI, offers a product that can be used to prepare not just a federal complaint, but all pleadings throughout the life-cycle of a federal case. Corporations with arbitration agreements with their workforce may be particularly vulnerable to these applications because they will make arbitration of employment claims even easier for disgruntled workers to initiate and because, in California and other jurisdictions, employers bear the full cost of arbitration.
Savvy law firms will lean into Generative AI in order to take advantage of emerging opportunities to offset any business lost to these products. Harvard Business Review writes that disruptive innovation permits the small firm with fewer resources to challenge established incumbents. (https://hbr.org/2015/12/what-is-disruptive-innovation) In other words, disruptive innovation is a risk to large established firms and an opportunity for smaller firms. In disruptive innovation theory, disruptions take place in markets that incumbents overlook. In the legal field, established white shoe national law firms typically adopt a strategy focused on serving its established market in Fortune 500 corporations. A foothold opportunity arises because these large incumbents overlook the needs of less-demanding clients. Insurance defense is an example of such a foothold market from the perspective of the mega white shoe firm whose rates are much higher than most carriers will pay. The smart firm will leverage Generative AI to dominate the foothold market and then use that position to compete with the incumbents. Success is not guaranteed. A successful disruptor must be nimble, both open to change and sufficiently well-funded to take advantage of innovations. This is one reason why start-ups can become successful disruptors. Their operations are fluid by design and often flush with venture capital funding. In the legal world, newer firms have an advantage over more established firms because they are more open to innovation and tend to have less rigid hierarchies.
Legal Considerations with Generative AI
The legal space has not yet caught up with AI innovation. The state bar of California recently asked the Committee on Professional Responsibility and Conduct to draft guidance for lawyers. Its report is expected by the end of 2023. Widely published stories reporting on “AI hallucinations” have resulted in an initial regulatory focus on the ethical obligation of competence under Rule 1.1. “A lawyer shall not intentionally, recklessly, with gross negligence, or repeatedly fail to perform legal services with competence.” In case California attorneys failed to get the message, the state recently published Comment 1 to Rule 1.1. which explicitly provides: “[t]he duties set forth in this rule include the duty to keep abreast of the changes in the law and its practice, including the benefits and risks associated with relevant technology.” (Emphasis added). There is also justified concern that the use of a non-proprietary tool like ChatGPT may violate the duty of confidentiality Rule 3-100. It is not hard to imagine an attorney malpractice claim arising against a firm based on the allegation that one of its attorneys failed to meet the statutory mandate “[t]o maintain inviolate the confidence, and at every peril to himself or herself to preserve the secrets, of his or her client” by thoughtlessly pasting confidential client information into an unsecured web tool.
While the promise of AI review of thousands of pages of documents is attractive, such processes must be vetted to ensure compliance with ethical obligations for the production of ESI. Just as a 6th-grade teacher is not receptive to a “my dog ate it” excuse, neither will a Court be sympathetic to an attorney who can’t explain why her firm’s sophisticated AI failed to identify and produce relevant information or erroneously produced privileged or proprietary documents to the opposing party.
As there is very little relevant case law relating to the professional use of AI tools, it is unclear how courts will handle claims based on their misuse. However, it is clear that there is no safe harbor if AI inadvertently violates a rule of law or court. Every attorney using AI tools must assume that they will be held accountable for their conduct.
What is the responsible practice manager to do? First, firms should acknowledge that late adoption of AI tools might pose an existential risk to their practice. Russia’s Vladimir Putin reportedly claims that “the nation that leads in AI will be the ruler of the world.” The same may be true for the adoption of AI technology among law firms. AI-driven tools promise access to efficient research and analysis at a much higher-than-human level of quality and accuracy. An associate-hungry midsized firm can use these tools to scale without having to hire to meet a surge in demand. This strategy would allow the smaller and midsized firms to compete against large, national incumbents on billable rates without sacrificing profits. Moreover, predictive analytics tools that can accurately assess cases, and forecast legal outcomes based on historical data specific to a particular jurisdiction and judge, and eventually a specific jury pool, will give firms using these tools an edge in litigation. Additionally, there is a substantial reputational advantage to being seen by clients as an “innovator” and an equal reputational cost at being seen as “stagnant”.
Second, AI works best in collaboration with a professional. For example, an AI tool may be useful for generating a first draft of a deposition summary, but that draft should be fine-tuned by an attorney. Similarly, an AI tool may be useful for generating draft contracts, but the ultimate reconciliation between the language of the contract and the needs of the client requires the judgment of an attorney.
Finally, it’s vital to recognize that the associate in the next office probably already has AI tools running. The prudent law firm will draft policies and provide training to these associates to mitigate risks and maximize benefits. The best practice is to bring both partners and associates into the development of these policies to encourage adoption and understanding throughout the firm’s hierarchy.