Generative AI marks a turning point in legal practice

The result has been a surge of experimentation and investment in AI for legal services. What was once a technology on the periphery of legal practice has moved to the mainstream. Lawyers are now using AI tools not only to search and review documents, but also to draft contracts, perform legal research, and even generate first-draft briefs. In early 2023, global law firm Allen & Overy became the first major firm to publicly announce a partnership with an AI startup, deploying a GPT-4 based chatbot named Harvey to help its 3,500+ lawyers draft documents and conduct research. Around the same time, Big Four firm PricewaterhouseCoopers (PwC) gave 4,000 legal staff access to Harvey for tasks like contract analysis, due diligence, and regulatory compliance, while emphasizing the AI “will not replace lawyers” or give independent legal advice. These high-profile moves signaled that generative AI in law had arrived, and that even historically cautious legal institutions were now embracing cutting-edge AI tools.

Rapid adoption of AI in the legal profession

Over the past two years, AI adoption in the legal sector has accelerated at an unprecedented pace. In the United States, the American Bar Association’s 2024 Legal Technology Survey found that 30% of law firms were using some form of AI, up from just 11% the year before. This nearly threefold jump in one year reflects how quickly generative AI has moved from novelty to normalcy. Larger firms are leading the charge. Among firms with over 100 attorneys, 46% report currently using AI tools, versus only 16% a year prior. But even solo practitioners and small firms are catching up. About 18% of solos use AI now, compared to effectively zero two years earlier. An additional 15% of firms not yet using AI said they are seriously considering AI purchases, suggesting the next year will bring even broader adoption.

This rapid uptake is a global phenomenon. A 2025 survey of 2,275 legal professionals and executives across 50+ countries found 72% view AI as a force for good in the profession, and 80% believe AI will have a “high or transformational” impact on their work within five years. Crucially, more than half (53%) of respondents worldwide say their organizations are already realizing returns on their AI investments, underscoring that AI in law is not just hype but is delivering tangible value. Law firms have moved from small pilot projects to firm-wide deployments. By early 2025, ChatGPT and similar generative AI became the most popular tools, and 52% of U.S. lawyers who use or plan to use AI have experimented with OpenAI’s chatbot. Even traditionally tech-shy firms are feeling competitive pressure. An April 2023 report noted 15,000 law firms were on the waiting list to try Harvey’s AI system.

“This is an arms race, and you don’t want to be the last law firm with these tools.”

Few in the industry want to be left behind as AI becomes standard.

Driving this adoption is the promise of greater efficiency and better client service. In surveys, a clear majority of attorneys cite time savings and productivity gains as the primary motivations for using AI. Routine legal work, like sifting through discovery documents or checking contract clauses, is being offloaded to AI, freeing up lawyers for higher-value activities. Clients, for their part, are pressuring firms to leverage technology if it means faster turnarounds and lower bills. In short, AI has swiftly moved from experimental to essential in the business of law.

Expanding applications across research, review, drafting, and more

The current generation of legal AI tools is being applied to a wide range of core lawyer tasks. Document review in litigation has seen extensive AI use for years (via technology-assisted review), and now newer AI systems are boosting that capability. Among legal professionals already using AI, 77% report using it to streamline document review. These tools can rapidly classify and prioritize large volumes of emails or PDFs in e-discovery, accomplishing in hours what might take a human team weeks. Similarly, legal research is being turbocharged by AI. Seventy-four percent of lawyers using AI say they employ it for research tasks. Instead of manually querying databases case by case, attorneys can ask an AI to find relevant precedents or summarize how a jurisdiction treats a particular legal issue. Major legal research providers have integrated AI assistants into their platforms. For example, Casetext’s CoCounsel, powered by GPT-4, can produce research memos or analyze deposition transcripts in minutes.

Another high-impact application is contract analysis and drafting. AI systems can quickly review contracts to flag risky provisions or missing clauses, a task once done by junior associates line by line. At PwC, the Harvey AI platform is used to analyze complex contracts and assist in drafting, helping ensure consistency and reducing human error in contract reviews. In corporate legal departments, AI is used to automate compliance checks by reading through regulations and company policies. Lawyers are also leveraging AI to summarize documents. Seventy-four percent of AI adopters use it to generate concise summaries of lengthy contracts or court filings. This has immediate productivity benefits, for instance allowing a quick synopsis of a 100-page brief before a meeting.

Even tasks like brief and memo drafting are seeing AI’s influence. Roughly 59% of lawyers using AI have employed it to generate first drafts of briefs, memos, or correspondence. While human lawyers must edit and refine the output, starting with an AI-generated draft can significantly cut down writing time. Some large firms have begun developing internal AI copilots tailored to their practice areas. For example, Holland & Knight is creating a tool to assist in editing financial agreements, and Baker McKenzie has pilots embedding AI in client services on a case-by-case basis. All of these applications point to AI becoming a “junior associate” of sorts, handling the labor-intensive groundwork of law practice. By automating the repetitive and data-intensive parts of legal work, AI allows attorneys to focus more on strategy, advocacy, and counsel.

Notably, these tools are not limited to back-office functions. They are increasingly part of client-facing work product. For example, a litigator might use AI to draft questions for a deposition or to analyze opposing counsel’s likely arguments. In transactional work, AI can assist in due diligence by reviewing thousands of documents in a merger, ensuring nothing important is overlooked. As one industry report summarized, AI is no longer theoretical in law. It is already embedded in a broad range of day-to-day workflows, from answering legal questions to aiding in knowledge management within firms. Every indication is that these applications will continue to expand as the technology improves.

Efficiency and ROI drive business impact

Law firms and legal departments are not adopting AI for novelty’s sake. They are doing so because it delivers concrete business benefits. Foremost among these is efficiency. AI-powered tools can dramatically reduce the time required for many legal tasks. One global survey estimated that current AI technologies have the potential to save lawyers nearly 240 hours per year, roughly 30 working days, by automating routine aspects of their jobs. In practice, this might mean getting a contract review done in a day instead of a week, or answering a legal question in minutes instead of hours of research. Such time savings directly translate into increased capacity and productivity. In the ABA’s 2024 tech survey, 54% of lawyers identified “saving time and increasing efficiency” as the number one benefit of AI, far ahead of any other benefit by a wide margin. With many firms still billing by the hour, efficiency might seem like a double-edged sword, but forward-looking firms recognize that greater efficiency can be parlayed into serving more clients, handling more cases, or offering alternative fee arrangements that attract business.

Indeed, AI may gradually help shift the business model of legal services. As repetitive work gets done faster, 43% of legal professionals anticipate a decline in the billable hour model over the next five years. Instead of billing 100 hours for a contract review that an AI can do in 10, firms may move toward value-based or flat-fee billing that better reflects expertise rather than effort. Lawyers can then reinvest their freed time into more strategic activities, deepening client relationships, developing novel legal arguments, or simply improving work-life balance. In one study, 42% of lawyers said they want to spend their liberated time on more engaging, judgment-based work as AI takes over the drudgery. Importantly, these efficiency gains need not come at the expense of quality. Many lawyers find that AI can increase the quality and consistency of work output by catching mistakes or standardizing best practices. For example, AI-assisted contract drafting can include the latest preferred clauses and flag any deviations from a company’s playbook, reducing the chance of human oversight errors.

The business impact is also evident in the return on investment. More than half of legal organizations using AI have already seen a positive ROI from their AI spend. This return can come from cost savings (for example, doing the same work with fewer personnel hours), from the ability to take on more matters, or from competitive advantages in winning clients. Early adopters are marketing their AI capabilities as a feature, advertising faster turnaround or more thorough analysis. In a client-driven industry, being able to say your firm uses state-of-the-art AI for efficiency can be a selling point. On the flip side, some corporate clients are beginning to expect their outside counsel to leverage technology to control costs. All these factors create a strong incentive for firms to invest in legal AI, not just as a tech experiment but as a core part of business strategy.

As one law firm innovation officer noted, “we expect [AI] to be a big deal, and we will use it,” capturing a growing consensus that AI is integral to the future of legal service delivery.

Investment and innovation in the legal AI ecosystem

The widespread interest in AI has unleashed a wave of investment and innovation in legal tech. Established legal technology companies and upstart ventures alike are racing to incorporate the latest AI into their products. A notable development was Thomson Reuters, one of the largest legal research and software providers, acquiring the AI startup Casetext in mid-2023 for $650 million in cash. Casetext was a pioneer in applying GPT-4 to law via its CoCounsel platform, and boasted a user base of over 10,000 law firms and legal departments. Thomson Reuters’ CEO described the deal as “another step towards bringing generative AI solutions to customers,” underscoring how critical the technology is for the future of legal research services. Around the same time, LexisNexis (Thomson Reuters’ chief competitor) launched Lexis+ AI, partnering with OpenAI to embed GPT capabilities into its research and drafting tools. These moves by legal information giants signal that AI features will soon be deeply integrated into the primary tools lawyers use every day, from research databases to word processors.

Venture capital is also flowing into legal AI startups. Harvey, the generative AI startup featured in Allen & Overy’s and PwC’s initiatives, raised a new $21 million funding round in April 2023 led by Sequoia Capital. The demand was evident. Harvey’s founders said more than 15,000 law firms had signed up on a waiting list to try its AI co-pilot. Other startups are targeting niche use-cases, for example AI tools specifically for contract review (such as Luminance and Kira Systems), for predicting litigation outcomes, or for automated compliance monitoring. The legal tech market has seen a boom in “AI-powered” solutions, and while not all will survive, the experimentation is rapidly expanding what’s possible. Law firms themselves are participating in this innovation wave. Several large firms have set up internal incubators or innovation teams to develop proprietary AI applications tailored to their practices. Firms are collaborating with tech companies and even with each other in some cases to train AI on unique datasets like brief banks or past deal documents, creating custom tools that give them an edge in efficiency or insight.

One significant aspect of this ecosystem is the partnership between law firms and tech providers. Big law firms that historically built bespoke software now often opt to partner with specialized AI startups. For example, dozens of major firms (DLA Piper, Orrick, Reed Smith, and others) in 2023 announced deals to adopt CoCounsel or similar GPT-4 based products for firm-wide use. These collaborations help AI developers refine their tools with real-world legal feedback, while giving firms early access to cutting-edge capabilities. The result is a virtuous cycle driving AI improvements. We are also seeing cross-pollination from other industries. For instance, accounting and consulting firms like PwC and Deloitte are leveraging their multidisciplinary expertise to build AI systems that handle tax and legal analysis together. All told, the legal AI sector in 2025 is a hive of activity, from big acquisitions and alliances to scrappy startups, all aimed at transforming how legal work gets done.

Challenges in accuracy, ethics, and regulation

For all its promise, the rise of AI in legal practice also brings a host of challenges and risks that the industry is grappling with. Foremost among these is the issue of accuracy and reliability. AI systems like large language models can produce outputs that sound convincing but are factually incorrect or legally flawed. A now-infamous episode in 2023 illustrated this danger. Two New York lawyers were sanctioned after they relied on ChatGPT to write a brief, only to discover that the AI had fabricated six case citations that did not exist. The judge fined the attorneys and emphasized that while using AI is not inherently improper, lawyers have a duty to ensure the accuracy of any AI-assisted work. This “hallucination” problem means that AI recommendations and drafts must be carefully reviewed by humans. In practice, firms are instituting strict quality control and validation processes around AI output, treating it like a junior employee whose work must be checked. According to one survey, 75% of lawyers identify accuracy as their top concern with AI, and reliability and data security rank close behind. The technology is improving, but a healthy skepticism remains essential whenever an AI tool is used for legal analysis.

Another challenge is maintaining client confidentiality and data security. Many AI tools, especially cloud-based ones, require uploading documents or data to external systems. This raises concerns about protecting sensitive client information. Law firms are understandably cautious about sending privileged documents to an AI service without ironclad assurances of security and privacy. In surveys, around half of attorneys cite data security worries as a key barrier to using AI. To address this, some firms are opting for on-premises AI solutions or ones that allow control over the data pipeline. Vendors are also developing “local” versions of AI models that can run behind a firm’s firewall. Ensuring compliance with privacy laws (like GDPR in Europe) when using AI is an ongoing consideration. Questions about who owns the output of an AI that was trained on proprietary legal data remain legally untested and add to caution in adoption.

Ethical and professional responsibility issues are also front and center. Lawyers must uphold duties of competence, confidentiality, and candor, among others, and these apply fully when using AI tools. In 2023, the American Bar Association issued guidance clarifying that a lawyer’s duty of competence includes understanding the “benefits and risks” of relevant technology like AI, and that lawyers must supervise AI just as they would a human assistant. Under existing ethics rules (ABA Model Rule 5.3), attorneys are responsible for non-lawyer assistance, which by definition now includes AI systems, meaning any output from an AI tool is ultimately the lawyer’s responsibility. Bar associations and law societies in multiple jurisdictions have warned that using AI doesn’t excuse negligence. A lawyer can’t defend an error by saying “the computer made the mistake.” Firms are accordingly developing internal policies on acceptable AI use, such as banning its use on confidential matters without client consent, or prohibiting AI from making final decisions. A nearly universal sentiment among practitioners is that human oversight is crucial. Ninety-six percent of lawyers in one global survey agreed that allowing AI to operate without human checks (for example, to represent clients in court) would go too far.

Regulators are also starting to pay attention to AI in the legal field. While no jurisdiction has yet enacted AI-specific law practice regulations, general AI regulations are emerging that will impact legal use. The European Union’s new AI Act, for example, classifies AI systems used in sensitive domains like law and justice as “high-risk,” subjecting them to strict requirements for transparency, quality, and human oversight. This means that AI tools used in European courts or legal decision-making could soon need to meet regulatory standards and undergo compliance assessments. In the U.S., state bar associations are studying AI’s implications, and at least one state supreme court has formed a commission on AI in law. Liability is another open issue. If an AI tool utilized by a law firm makes a critical error, the firm could face malpractice exposure, and questions may also arise about what responsibility, if any, falls on the software provider. Such questions remain largely hypothetical, but they loom larger as AI’s role grows. In short, the legal profession’s adoption of AI is being matched by a careful effort to draw ethical boundaries and safeguards. The consensus is that AI should remain a tool under human control, useful for improving service, but not a substitute for a lawyer’s judgment or accountability.

Implications for lawyers, clients, and the future of law

The integration of AI into legal work is beginning to reshape the skills and roles that lawyers need. Rather than rendering attorneys obsolete, a fear expressed early on, the evidence so far suggests that AI is transforming jobs, not eliminating them. An overwhelming 85% of professionals surveyed believe that AI adoption will require lawyers to take on new roles and learn new skills, rather than lose their jobs outright. One emerging expectation is that lawyers will need to become adept at working with AI, effectively acting as pilot or editor to the AI’s first draft. This means developing skills in prompt engineering (formulating the right questions for AI), in critically evaluating AI outputs, and in understanding the limitations of these tools. Many firms are already training staff on how to use AI-assisted research and drafting platforms. Law schools, too, are beginning to incorporate AI literacy into curricula, recognizing that tomorrow’s lawyers must know how to leverage these technologies responsibly.

We can also expect new specialist roles to arise. In surveys, legal professionals anticipated increased demand for positions like AI legal technologists, AI implementation managers, or legal data scientists to help firms deploy and monitor AI systems. Some large law firms have created internal teams focused on AI governance, ensuring the firm’s use of AI meets ethical and quality standards. On the client side, corporate legal departments may prefer hiring firms that demonstrably use AI to be more efficient, which could stratify the market between tech-forward firms and laggards. Clients are also exploring AI for their own in-house tasks, such as automatically reviewing incoming contracts or monitoring regulatory changes. This could potentially shift some work away from outside counsel if corporate legal can handle more routine matters with AI in-house. Conversely, if AI enables law firms to lower costs, it could improve access to legal services for currently underserved markets, by making it economical to serve individuals or small businesses with lower budgets. Enhanced productivity might allow more affordable legal advice or fixed-fee offerings, something 20% of lawyers in one survey hoped would be a future outcome of AI adoption (for example, improving access to justice).

For the broader legal system, the advent of AI may prompt updates in regulations and even in law itself. Courts might develop rules on when AI-generated material is admissible or how to cite AI outputs. Intellectual property law will wrestle with questions of AI-generated legal content ownership. Professional standards will likely evolve. For example, bar exams and continuing education may incorporate more technology components. We may also see changes in how law firms are structured and bill. If AI reduces the need for armies of junior associates to grind through documents, firms might become leaner or focus on hiring talent with complementary skills (like data analysis or programming). Billing models could shift toward project-based fees or subscriptions rather than hourly billing, as efficiency increases. Notably, nearly 45% of lawyers believe AI will become mainstream in law within three years.

In sum, the trajectory points to a legal landscape where AI is embedded in the fabric of legal work. The roles of lawyers will adapt, with more emphasis on strategic thinking, advisory, and ethical oversight, and less on brute-force research or routine drafting. Firms that successfully blend AI’s strengths with human expertise are likely to outperform those that don’t. The fundamental mission of legal professionals, solving client problems, advocating justice, providing counsel, will remain. But how those services are delivered is poised to fundamentally change, with implications ranging from training and hiring to client relationships and economics.

A transformative but human-centered future

The rise of AI in the legal profession represents one of the most significant shifts in how legal services are delivered in decades. Unlike past innovations that were confined to narrow tasks, today’s AI has the versatility and intelligence to influence virtually every area of practice, from litigation to corporate deals to regulatory compliance. The evidence so far, rapid adoption rates, substantial efficiency gains, and improving AI capabilities, points to a future in which AI is an integral partner in legal work. Early movers are already seeing competitive advantages, and even traditionally conservative institutions are acknowledging that mastery of AI will be a hallmark of the successful law firm or legal department. As one law firm leader noted, nobody wants to be the “dinosaur” that clings to paper and precedent while rivals harness intelligent automation.

Yet, for all the technological transformation, the consensus in the industry is that law will remain a fundamentally human-centric endeavor. AI can draft a competent contract or brief, but it takes a seasoned lawyer to craft a winning legal strategy, exercise judgment, and ethically stand up for a client’s interests. The most effective uses of AI treat it as an augmenting tool, a way to achieve better outcomes more efficiently, rather than a replacement for human lawyers. Indeed, regulatory and ethical frameworks are being updated to ensure AI is used responsibly, under human supervision and with appropriate safeguards. The successful law firms of the future will be those that pair technological prowess with unwavering professional standards.

In the final analysis, AI’s growing role in the legal profession is both transformative and inevitable. It promises a world where lawyers can devote more time to complex problem-solving and client counsel, while mundane tasks are handled by tireless algorithms. It may also democratize legal services, as AI lowers costs and enables new delivery models. Realizing these benefits depends on care in implementation, including verification of AI outputs, vigilance about biases or errors, and an adherence to the core values of justice and integrity. The legal profession has begun this journey in earnest. The coming years will determine how seamlessly AI is woven into the practice of law, not as a usurper of the lawyer’s role, but as a powerful new tool in the lawyer’s toolkit, driving a more efficient, accessible, and innovative legal system for all stakeholders.