ANALYSIS: AI in Law Firms: 2024 Predictions; 2025 Perceptions

When generative AI first made its debut, it sparked bold predictions of dramatic change to the legal practice. Efficiency at law firms would soar, mundane tasks would be a thing of the past, and lawyers would finally be free to focus on actual lawyering.

This broad sense of optimism was reflected in a 2024 Bloomberg Law survey in which law firm lawyers were asked how they anticipated generative AI would impact workloads and billing practices in the profession. Respondents expressed high expectations and predicted that generative AI would deliver significant improvements in efficiency, raise both the quality and quantity of legal work, and bring about wide adoption of alternative billing models.

Fast forward one year, and the reality of AI’s impact has fallen short of early expectations. In Bloomberg Law’s 2025 State of Practice Survey, law firm lawyers were asked for their perceptions on how generative AI has actually impacted those same legal workloads and billing practices in the legal industry.

While a majority of attorneys have used AI in some way for work, most also have observed limited effects from AI in general during this period and reported smaller-than-expected changes in every workload and operational category of the survey.

Fewer Changes Than Expected

Lawyers were most optimistic about AI’s potential to increase the number of processes or workflows automated. In 2024, 75% of respondents predicted an improvement in this category. Although this turned out to be the category in which the most respondents have seen an increase in 2025, the scope was far more modest, with 37% of lawyers reporting that AI has increased the number of automated processes and workflows.

A similar pattern played out across other areas. In each category, the percentage of surveyed lawyers who observed AI-driven increases in 2025 turned out to be much smaller than those who anticipated such changes in 2024.

For example, increases in the amount of time spent on both billable and nonbillable work, both for attorneys and support staff, were predicted by respondents in 2024 at a higher rate than what the 2025 results actually show.

When it comes to billing in particular, nearly four in 10 law firm attorneys (39%) expected AI to accelerate adoption of alternative fee arrangements, anticipating shifts to long-standing billing practices. This, however, proved to be one of the largest expectation/reality gaps between 2024 predictions and 2025 perceptions, with only 9% of respondents reporting increases in law firm adoption of AFAs.

Explaining the Gap

While several factors likely contributed to the smaller-than-expected impact of AI on law firms’ operations, the simplest and most compelling explanation may be that it’s still too early to tell.

We know that both in-house and law firm attorneys are experimenting with generative AI in some capacity, but the full effects have yet to materialize. The fact is, across all categories, the vast majority of respondents in 2025 reported that they see “no change” as a result of AI’s involvement in the legal industry. And the number of those who are noticing decreases in these areas has turned out to be smaller than expected as well.

Essentially, these results reinforce the idea that the true impact of AI on legal workflows hasn’t yet to fully take shape and it may be too soon to gauge. This uncertainty is compounded by the fact that AI tools haven’t yet been widely adopted for more complex, high-stakes legal work such as securities filings or intricate litigation. Until broader implementation reaches these substantive areas, the true impact of AI on legal practice will remain a matter of wait and see.

The gap between expectations and reality can also be attributed to the inherently complex nature of legal work coupled with the profession’s traditional risk aversion. While the AI boom presents significant opportunities, it also introduces a range of ethical, security, and legal risks for law firms. These include threats to client data and confidentiality (especially when data is stored with third-party providers), privacy concerns, cybersecurity vulnerabilities, intellectual property challenges, and the potential for data bias inherent in many AI technologies.
Lawyers’ professional duty of technical competence in this context is particularly challenging. It requires a thorough understanding of how AI tools function, evaluating the accuracy and fairness of their outputs, staying informed about evolving risks and benefits, and clearly communicating these aspects to clients.

Meeting this standard has become increasingly demanding due to the fragmented regulatory landscape and the proliferation of various requirements and guidelines. Executive orders, court rulings, and standing orders, state-level rules, international standards and guidelines have created a patchwork of obligations and expectations around AI use and have likely impacted AI deployment in law firms.

Navigating this constantly evolving legal terrain—combined with dealing with the reputational harm, sanctions, and fines that come from lawyers’ improper AI use and hallucinated citations—is also likely fueling law firms’ hesitancy to adopt AI tools wholesale. Rather than risking ethical missteps or regulatory violations, firms appear to be opting for a more cautious approach, delaying broader implementation until clearer and more consistent regulatory frameworks emerge.

Even for firms that are embracing AI tools, it is conceivable that transitioning from legacy systems, getting security credentials in place, setting up training programs, and handling a host of other integration challenges may be making implementation more complex than anticipated.

Alignment in the Future?

The expectations of 2024—fueled by the novelty of AI capabilities, vendor hype, and public enthusiasm—haven’t fully aligned with the operational realities of legal practice in 2025, contributing to the lower-than-anticipated pace of AI impact. This, however, is certainly not a rejection of AI, but rather a reflection of the persistent complexity of legal work and the cultural inertia of a risk-averse industry.

Most lawyers are using generative AI in their legal work in some capacity. Foundational shifts are very much underway. And as AI technologies mature and law firms build more robust frameworks to manage risk, ethics, and integration, we are likely to see an impact closer to what lawyers predicted.

The new report, Artificial Intelligence: The Impact on the Legal Industry, is available to subscribers here. Non-subscribers can click here to download the report.

In previous articles in this series: Bloomberg Law Legal Analyst Eleanor Tyler’s July 29 article examined AI slop in litigation and possible remedies. Bloomberg Law Legal Analysts Janet Chanchal’s and Linda Masina’s Aug. 13 piece looked at AI washing and how the legal profession can curb it. Stephanie-Solange Campbell’s piece on Colorado’s new AI law will publish next on Aug. 18.

Bloomberg Law subscribers can find related content on our In Focus: Artificial Intelligence resource and our AI Legal Issues Toolkit.

If you’re reading this on the Bloomberg Terminal, please run BLAW OUT in order to access the hyperlinked content, or click here to view the web version of this article.