eDiscovery, legal research and legal memo creation - ready to be sent to your counterparty? Get it done in a heartbeat with AI. (Get started for free)
One of the most time-consuming aspects of legal work is the discovery process. Discovery involves going through mountains of documents, emails, and other materials to find key evidence relevant to a case. Traditionally, armies of junior lawyers and paralegals have had to spend countless hours reviewing and analyzing documents by hand. However, the rise of AI is automating significant parts of this process and leading to huge time and cost savings.
E-discovery software powered by machine learning algorithms can review documents and automatically tag them by predicted relevance, pull out key names and dates, and summarize long emails and memos. This allows lawyers to quickly focus their attention on the most important materials rather than wading through oceans of texts. According to Casey Flaherty, corporate counsel at Kia Motors, manual review costs a minimum of $40,000 per gigabyte of data. With AI software, overall discovery costs can be reduced by up to 30-40%.
Leading legal tech companies like Everlaw, DISCO, and Logikcull are at the forefront of using AI for discovery. For example, Logikcull's Discovery Assistant uses optical character recognition and natural language processing to index files, extract key information, and sort documents into categories. Users can then search across millions of documents in seconds. Everlaw's platform goes even further by predicting document relevance and identifying concepts and entities within texts. This allows lawyers to prioritize the most important evidence and pinpoint connections.
Many major law firms have adopted these AI tools. Dentons, one of the world's largest law firms, uses DISCO for faster document review. O'Melveny & Myers has an entire E-Discovery Excellence group that leverages AI and advanced analytics. Smaller firms are also jumping on board. Mike Hamilton, a partner at Nashville firm Dickinson Wright, shared that "the technology has taken a once lengthy, manual task and reduced it to minutes."
While AI simplifies document review, human guidance is still essential. Lawyers must still determine key search terms and strategically sample results to ensure accuracy. According to Jake Frazier at legal tech company Intapp, "Technology will be most effective if the focus remains on using AI to augment lawyers" skills, rather than replace them." Still, AI automation is allowing lawyers at all levels of experience to spend less time on routine discovery work and more time crafting legal strategy.
One of the pillars of legal work is researching past laws, cases, and rulings to find precedents and build arguments. Traditionally, this has involved lawyers spending hours painstakingly combing through legal databases like Westlaw and LexisNexis. But now, AI tools are quickly analyzing millions of legal documents and extracting key insights. This allows lawyers to get to the information they need faster than ever before.
According to Bob Ambrogi, an expert on legal tech, "The greatest promise of AI in the practice of law is to expand access to justice by making legal research faster, better, and cheaper." For example, services like Casetext CARA and ROSS Intelligence leverage natural language processing algorithms to quickly analyze briefs, contracts, and case law. They can then pull out the most legally relevant passages and pieces of evidence tailored to the specifics of a lawyer"s case.
Some systems like Lex Machina even focus on a specific practice area. Lex Machina mines litigation data and outcomes to provide insights for IP lawyers on judges, lawyers, parties, and more. This allows IP attorneys to craft targeted legal strategies based on data-driven insights. According to Brian Howard, Lex Machina"s legal data scientist, "Legal judgment is improved by understanding details of specific judges, lawyers, parties, and the like."
Firms leveraging these AI research tools have reported massive time savings that allow them to be more competitive. As Stacy Stern, Managing Partner of Underhill Kesselring LLP commented about using Casetext CARA, "I can litigate with the big dogs now." She shared that an analysis that would have taken three days now only takes her three hours, allowing her to compete with larger firms.
That said, AI research tools are not a magic bullet. As stated by Reynaldo Aligada, Head of Legal Tech and Innovation at a top UK law firm, "Technology like this won"t replace lawyers, but it will help them refocus their skills on tasks that matter and clients value most." The tools serve to augment human expertise, not replace lawyers entirely. As with discovery, human guidance is key. But by automating the tedious parts of research, AI enables lawyers to concentrate their skills on higher-value analysis and strategy.
The power of AI-driven research is only just being tapped. As algorithms grow more advanced, lawyers can expect even more comprehensive insights and time savings. While AI won"t be writing legal briefs from scratch anytime soon, by supercharging legal research, it is undoubtedly transforming how lawyers find and leverage information. This will only continue accelerating in the years to come as legal tech becomes more mainstream.
One frontier of legal AI that holds great promise is using natural language generation to automatically create legal documents. While AI is not yet advanced enough to fully replace human drafting skills, it can help automate the creation of standardized legal documents and provide a starting point for attorneys to work from. This has the potential to significantly reduce workload and costs for firms.
Many legal tasks like contracts, briefs, and memos involve standard formats and repeatable components. Natural language generation uses vast datasets and deep learning to essentially "study" how effective legal documents are written. The AI can then generate new documents by pulling key phrases and paragraphs from its database and adapting them to new facts and contexts.
Leading companies pioneering this technology include LawGeex, LegalSifter, and CaseFix. LawGeex"s AI reviews contracts and suggests missing clauses, legal language, and potential risks. LegalSifter"s Writer tool generates first drafts of transactional documents like non-disclosure agreements. CaseFix goes even further by generating personalized plaintiff demands based on the unique facts of a case.
Am Law 200 law firm Davis Wright Tremaine has tested CaseFix for automating early demand letters with positive results. As partner Ronnie Fischer shared, "The AI took a tremendous amount of work off the attorney"s plate. It reduced something that would take hours to create to just minutes of review." The Partner estimated a cost savings of 20-30% for each demand letter.
According to Richard Tromans, founder of legal AI website Artificial Lawyer, the benefits go beyond time savings: "The drafts also help junior lawyers improve their legal writing skills by giving them high-quality documents to learn from." This on-the-job training makes associates better lawyers.
That said, Tromans notes that a lawyer must still be in the loop: "The ideal process has a lawyer reviewing each draft document before sending to the client, using it as a starting point, not an end product." The AI generates the first draft, and attorneys polish, edit, and finalize.
Striking the right balance of automation versus human oversight is key. As Prashanth Yadavgiri, VP of Product at Thomson Reuters Contract Express, advised: "The objective should be maximizing the time lawyers spend exercising their judgment, not mechanically editing contract documents." AI document creation aims to free lawyers to focus on high-value strategic tasks, not fully replace their skills.
The ability to make data-driven predictions is transforming many industries, and the legal field is no exception. Predictive analytics leverages massive datasets and machine learning algorithms to forecast outcomes and risks for court cases, contracts, and more. This emerging technology holds both great promise and potential pitfalls.
On the promise side, predictive analytics can help attorneys better evaluate cases and give clients realistic assessments earlier on. For example, companies like Lex Machina and Premonition mine millions of legal documents and outcomes to generate insights on judges, opposing counsel, and likely case resolutions. Lex Machina boosted one firm"s win rate from 30% to 80% by revealing that a certain judge harshly penalized companies violating injunctions. Having an accurate prediction early allows lawyers to set client expectations and craft an effective legal strategy.
Premonition even makes predictions at the level of individual lawyers by analyzing millions of court records and filings. As CEO Guy Kurlandski explained, it assesses things like "how many cases a lawyer has, types of motion filed, cases won or lost and so on." This data helps clients pick the best legal representation. According to general counsel Alan Bryan, "Knowing counsel"s record in similar cases is highly valuable in deciding who to hire."
On the concerning side, predictive analytics raises risks of perpetuating biases, breaching privacy, and discouraging settlement. The algorithms are only as good as the data they are based on. Critics like Frank Pasquale argue that if past court decisions contain racial or gender bias, predictive algorithms will propagate those same injustices. Privacy is also a concern as insights may be derived from personal information without people's knowledge or consent.
There are also concerns that if both sides can predict outcomes, they will become overconfident in success and unwilling to compromise. As stated by Daniel Katz, associate dean at Chicago-Kent College of Law, "Parties might get unreasonable expectations about their chances if they knew what the models said." Settlements could decrease, causing disadvantages for the court system.
The rise of artificial intelligence is fueling speculation about the future of the legal profession. With AI performing tasks like legal research, document review, and contract analysis, many wonder if robot lawyers could one day fully replace human ones. While true automation is still years away, AI is impacting how lawyers work and the skills in demand.
Complete replacement is unlikely in the near future, but AI tools allow lawyers to be more efficient by taking over routine and repetitive tasks. As stated by Mike Hamilton, a partner at Dickinson Wright, "Where AI really helps is doing away with the grunt work that no one likes to do." At Hamilton"s firm, AI does the first-pass document review, reducing 360 hours of work to just 3. As Hamilton explained, this allows lawyers to "get to the good stuff faster." The role becomes more advisory as AI handles tedious minutiae.
That said, human skills like emotional intelligence, creativity, and judgement are still vital. Silicon Valley pioneer Oren Etzioni stresses that "Lawyers worry about ethics, legality, morality. Software doesn"t do that." Nuanced legal strategy requires human critical thinking and ethics that AI lacks. High-stakes cases and negotiations need advisors who understand client priorities and can gain trust through empathy. As stated by Reynaldo Aligada, Head of Legal Tech and Innovation at a top UK law firm, "There are subtleties in engaging with clients that require emotional intelligence. That's very hard for technology to replicate." Human lawyers provide value AI tools cannot.
The most likely future is AI and lawyers working together to complement each other"s strengths. As explained by Intapp VP Michelle Furlough, "The goal should be using tech to enhance human skills and relationships, not remove humans entirely." AI excels at tasks like discovery and research while humans handle strategy, ethics, and client interactions. Stanford Law Professor Daniel Katz predicts that "the lawyer of the future will leverage technology to serve more clients better." Specialized AI platforms will allow lawyers to take on more clients and focus expertise on high-value work.
As AI systems grow more advanced, a major application will be AI assistants that are integrated directly into legal teams to augment human capabilities. While not ready to fully replace lawyers, these expert systems hold promise for supercharging how legal work gets done.
A major benefit of AI assistants is providing constant support for common legal tasks. Junior lawyers currently spend many hours on routine research and document review that divert them from higher-value work. AI tools embedded within firms could handle a large volume of these repetitive tasks and enable lawyers to focus on strategy and client needs.
For example, LawGeex offers an AI assistant named LISA (Legal Intelligent Self-Service Assistant) that handles contract review. LISA scans contracts in seconds to identify missing terms, flag risks, and suggest edits lawyers may overlook. This constant support helps lawyers work faster and avoid mistakes. As noted by LawGeex CEO Noory Bechor, "AI technology enables attorneys to provide a higher quality of legal services at a fraction of the time and cost."
Other companies like Ross Intelligence and Luminance offer virtual assistants focused on legal research. These tools use natural language processing to answer common legal questions posed in plain English. Having an "expert on call" boosts productivity and on-the-job learning. As Deborah Berecz, head of legal operations at Goldman Sachs, explained of their AI assistant, it allows lawyers to "get "knnnnowledge" at their fingertips" saving time while building skills.
AI assistants also offer major advantages in terms of scale and consistency. While human associates may tire or make mistakes, AI tools can provide continual support without fatigue or errors. Luminance highlights that their platform "ensures every document gets the same detailed level of analysis...without fail." Such consistency and tirelessness provides value human staff cannot match.
That said, AI assistants function best alongside human professionals, not in place of them. As noted by Reynaldo Aligada, head of legal tech at a top UK law firm, the tools serve to "enhance humans, not replace them." Human oversight ensures AI does not run amok. Lawyers also provide the empathy, creativity and contextual judgment that AI lacks. Partnership allows each to play to their strengths.
A major frontier in legal AI is developing systems capable of replicating the complex legal reasoning and argumentation that human lawyers employ. Rather than just searching databases and automating simple tasks, truly intelligent legal AI needs to be able to construct arguments, weigh competing interpretations, and provide nuanced legal analysis like an attorney. This has prompted increased focus on training AI models on legal reasoning.
According to Emily Foges, CEO of legal tech company Luminance, "The real transformative benefits will come when contracts can be seamlessly analyzed for the merits of arguments, not just their existence." Moving beyond keywords to evaluate the strength of arguments requires abilities like logic, critical thinking, and understanding context and exceptions. Teaching these human-like skills to AI is challenging but critical for advanced legal applications.
Some researchers are trying to directly train AI systems on the multifaceted arguments in court cases. For example, researchers at the University College London compiled a dataset of hundreds of real court judgments annotated with argument components like claims, premises, and reasoning types. They used this corpus to train a model called ARG-IL to extract interconnected arguments from legal texts. While still basic, such models are a first step toward AI capable of construing legal logic.
Other efforts have focused on training AI to argue persuasively. At Duke University, Professor Anne Klinefelter led development of an algorithm to generate legal arguments both "for" and "against" hypothetical scenarios. The system learns persuasive techniques like using emotion, precedent, and policy arguments. According to Klinefelter, "This research is teaching legal AI how to use techniques of advocacy and reasoning." Exposure to persuasive arguments teaches the AI how experienced lawyers construct narratives around the law.
Researchers are also exploring how transformers like GPT-3 can be adapted to legal reasoning when given the right training data. As explained by Christian Lang, CTO of legal AI company Aletha, "With fine-tuning on legal corpora, we have seen transformers make counterarguments, weigh strengths of precedents, and compare the applicability of past cases." By learning from quality examples, transformers show surprising ability to engage in lawyer-like analysis.