It’s time for legal AI strategies to catch up with legal AI adoption
As lawyers, we don’t usually aspire to be rapid tech adopters. When it comes to new technologies, it’s usually our role to point out the risks, control unauthorized use, and urge organizations to consider consequences before ploughing ahead.
When it comes to AI, however, something very different is playing out in our industry. The last year has seen a rapid transformation in the way that legal work is done. In 2025, DeepL surveyed 1,000 legal professionals in the US, and 77% told us that their organizations increased their AI spending in the previous year and 96% are now using AI at work, with 47% describing AI as an essential. Those surveyed included lawyers from established firms, in-house legal counsel and public sector teams. We found the pace of AI adoption accelerating across all of them.
From pressing the brake to pressing the accelerator on AI
Far from proceeding cautiously or seeking to apply the brakes, the law has become a test case for just how quickly AI can shift assumptions, accelerate productivity, re-imagine workflows and re-engineer business models. Perhaps surprisingly, it has also become a case study for how tempting it is to start sidestepping guidelines and compromising compliance, once such transformations take hold. For many people in our profession, the most striking finding of the DeepL survey might be the fact that 71% of legal professionals admit to using AI tools without formal approval, and half of those doing so say they use those tools not just now and then, but frequently.
This rapid spread of shadow AI feels like a shock given the importance of security, privacy and compliance in legal work. It’s certainly not a concern that’s confined to the US. In my conversations with leading law firms around the world, I find the same issues raised, again and again. At a London roundtable organized by The Lawyer magazine, guests spoke about how their colleagues needed to, “realize what they’re actually interacting with when they interact with AI.” There was a common consensus that safeguards were lacking and that unsanctioned AI use didn’t just risk data breaches, but undermined the quality of legal advice. When AI generates hallucinations in contracts or legal arguments, an entire firm’s credibility could be swiftly undermined.
Rather than reversing adoption, let’s get strategy up to speed
The response to such concerns can’t be to try to reverse the transformation that has swept across the legal profession. The real cause of this transformation spinning out of control is that AI strategy hasn’t moved fast enough to keep up. The best solution is to fix that. It’s the key to integrating AI effectively into legal teams, elevating their expertise and credibility rather than undermining it, and turning breakneck AI adoption into an AI strategy with credible compliance, real goals and a real Return on Investment (ROI).
Approved AI platforms that can keep pace with innovation
Building a more effective AI strategy starts with acknowledging where the current disconnect comes from. In our survey, the most common reasons for using unauthorized AI tools were pressure to deliver work faster, and a lack of functionality among approved AI tools. Those pressures often come from clients demanding greater efficiencies (35% of legal professionals cite pressure to deliver work faster as the reason for using shadow AI), but they can also come from senior leadership navigating a transformation of business models and client relationships. In fact, 30% of respondents say they’ve used unauthorized tools that were recommended by their managers or senior colleagues.
One of the findings that stood out to me in our survey is that a third of legal organizations are already considering shifting to fixed-fee or value-based billing as a result of AI, and just as many have already revised their internal cost assumptions. Over 40% say their teams are now delivering more work with the same headcount. Once those gains are locked into the strategy, and into client expectations, turning back the clock becomes impractical and uncompetitive. What organizations need are approved tools that can deliver the compliance, quality and security they need, while still delivering the gains that they, and their clients, are counting on.
Once we acknowledge that AI adoption is as essential as legal professionals tell us it is, the focus shifts to finding a better strategy for that adoption. Inviting employees to explore the AI landscape for themselves sounds progressive and innovative, but it’s fundamentally flawed as an approach. What legal professionals need is clear guidance and direction. Leaders of firms and in-house teams need to direct their people towards AI tools that meet their requirements on security, accuracy and compliance. Even better, they need to choose AI partners that understand those requirements and will continue to deliver innovative, new solutions that align with them.
Combining AI innovation with innovation on compliance
This is where DeepL comes in. We work with a lot of legal teams, because language is one of the most valuable applications of AI for them. In fact, 85% of the legal professionals we surveyed say that translation issues have affected their work, which is why 45% of legal organizations have approved multilingual AI tools. However, that’s not the only reason that DeepL has a central role to play.
For a legal team, it’s not enough for an AI translation tool just to translate words accurately. They need a commitment to innovation in security and compliance that’s as strong as the commitment to product development. They need AI that has a deeper, contextual understanding of language in the legal context to ensure precision in translation and quality in output. And they need tools that can be customized to reflect the expert knowledge of a firm, company and team, and elevate that expertise.
We’ve developed DeepL as an AI platform that’s specifically designed to meet these requirements. It’s why we comply with the latest security standards for the world’s most sensitive industries, and why we continually add new features, such as Bring Your Own Key encryption, Multi-factor Authentication (MFA) and advanced access controls, which embed compliance without compromising productivity. DeepL enables AI translations that are rated the most accurate by language experts, and which are customized to each organization.
The value of a secure, compliant and yet agile AI strategy goes far beyond heading off the risk of shadow AI. Rather than being hostage to the pressure of AI adoption, organizations are able to make confident, strategic investments which have clear outcomes, and for which they can plan and calculate a real return. An AI strategy shouldn’t just be a rushed piece of risk mitigation. With the right tools, platforms and partners, it’s an opportunity to elevate the work that you do, transform your impact as a team, and unlock growth opportunities. And with the right partner, it’s an investment that will keep delivering for you.