A new report by Morae, prepared with Coleman Parkes based on 850 surveys of senior legal professionals in the United States, the United Kingdom, Australia and the Middle East – the sample does not include firms from Latin America – confirms that 67% of lawyers and general counsel believe that the cost of verifying the results of artificial intelligence already exceeds the benefits that the same technology promised to deliver.
This is the report’s central finding, but not the only one. The industry has adopted AI at breakneck speed without addressing the underlying problems that determine whether that adoption actually works.
The first is trust. 46% of organizations already have AI integrated or embedded in their legal processes, and another 31% use it regularly in defined workflows. But only 33% have full or mostly confidence in the results it produces, and 48%—almost half the sector—admit that they manually correct those results before they can use them, in most cases substantially.
Moreover, this correction is often done without sufficient preparation. Only 29% of professionals feel equipped to monitor AI results, 26% to question them, 22% to defend them before a client or regulator, and a mere 19% to consistently integrate them into their workflow. Nearly half of the sector intervenes in AI results without a structured framework to understand what they are correcting—and what they might be overlooking.
The second problem is data, and it explains a good part of the first. Only 26% of legal leaders feel confident that their organization’s information is ready for AI to work with effectively.
The reason is structural: 44% admit to having multiple internal systems without any centralized control, 39% lack a complete inventory of their legal information, and another 39% are burdened with legacy data that was never classified or cleaned. An AI built on this foundation doesn’t solve the problem; it inherits it and returns it as results that must be reviewed one by one.
The third problem is governance. Nearly a third of firms lack any formal framework to regulate AI-generated outcomes. And among the 70% that do have one, only 32% consider that policy to be fully implemented—the rest admit it exists more on paper than in practice. Having a policy and consistently applying it are, according to Morae, two different things, and it is in this gap that the risk accumulates.
The three issues—trust, data, and governance—converge in the relationship between law firms and clients, where tension is already evident. 73% of legal professionals agree that clients should have full visibility into when and how AI is used in their cases, but only 21% of in-house legal departments describe that transparency as high or complete, and 41% rate it as limited or nonexistent. 80% of general counsel express concern about how external firms use their data in AI systems.
Law firms and clients, however, do not perceive the risk in the same way. Law firms are more alert to the possibility that their AI results could be challenged in litigation (29% compared to 16% of in-house teams), while in-house teams report weaker databases (40% compared to 32% of law firms) and feel less prepared to defend AI-assisted decisions to their own management.
Despite everything, investment continues. 58% of organizations plan to increase their spending on legal technology, AI, and information governance over the next two years. What’s revealing is that this investment is no longer concentrated solely on tools (65%), but is almost equally divided between information governance (48%) and AI governance (47%), even surpassing cybersecurity.
The impact on the workforce follows a similar logic. 33% of in-house departments are already seeing roles reinforced with a more analytical and supervisory focus, a figure expected to reach 52% in two or three years, while the creation of specialized AI governance positions would increase from the current 22% to a projected 41%.
The impact on billing models, however, is progressing more slowly: 56% expect minimal or no changes in the short term, although 39% do anticipate fee adjustments in the next 12 to 18 months, driven by pressure from the clients themselves.
The market, the report concludes, has understood that the gap between the promise of legal AI and its current performance is not a technology problem, but a problem of
implementation.
The report was prepared by Morae , a global provider of legal services and consulting, in collaboration with the research firm Coleman Parkes, between February and March 2026.
