The AI advantage: What’s real; what’s right and what’s risky for law firms today?

"AI won't just speed up the legal system - it will revolutionise it". Headlines like these are loud, and some of them present fascinating scenarios. Will AI-powered courts replace human judges for low-stakes civil disputes, delivering near-instant rulings based on vast legal precedent databases? Could algorithms analyse potential jurors' digital footprints, social media activity and associated psychological profiles to predict biases and select the most 'favourable' jury for a case?

But beyond the hype of speculation around when 'science fiction' will become 'science fact', it can be challenging for the legal profession to understand how AI is actually being used in the industry.  

We delve into the urgency surrounding the adoption of AI in the legal profession, the danger of investing without understanding or strategy, and why risk shouldn't be ignored in the race to modernise. 

Drinking from the fire hose

Understanding AI's implications for the legal profession can feel a lot like drinking from a firehose. There's an undeniable urgency and firms and legal departments are racing to adopt AI- tools, worrying they'll fall behind. Some with deep pockets buy and deploy AI-powered applications with a 'first-mover' mentality, driving the goal of leaping ahead of competitors. 

But without a thorough understanding of AI's capabilities and how these can be leveraged (safely) for legal use cases specifically, you risk adopting a fragmented approach at best. Before adoption, you must articulate precisely why and how AI will deliver value within your unique business, legal, and operational situation.

Do we really know what we're getting into?

The truth is that beyond the hype, it's very easy to get confused…

Take Generative AI (Gen AI) and Robotic Process Automation (RPA), for example. Many clients struggle to differentiate between Gen AI and RPA in the context of adoption efforts within legal departments. 

> Probabilistic vs deterministic insight

These two entirely different technologies promise efficiency but serve distinctly different purposes. Gen AI can draft contracts and provide likely negotiation points, summarise case law, and even forecast the likelihood of success in litigation.  It generates new content based on patterns it has "learned" via feedback and other processes.  When applied effectively, this technology will create real value for legal practitioners and clients alike. For the most part, because its outputs are probabilistic rather than deterministic, ongoing human oversight or engagement remains crucial. Defining the task it is supporting is vital. Is it a tool, an assistant, a peer or something else?  

> Rule-based automation vs meaningful insight

RPA, on the other hand, is a legal workhorse. RPA simulates how humans interact with digital systems and relies on the codification of activities to facilitate rule-based automation.  For instance, it can effectively manage data entry tasks such as invoice processing,  filing court documents and creating reports, all at lightning speed. It's structured, predictable and follows strict instructions without deviating. Unlike Gen AI, which simulates human-like reasoning, it simply executes predefined instructions to automate repetitive tasks - taking the 'boring' out of the job. 

Some may expect Gen AI to perform like RPA and flawlessly execute tasks they set with 100% accuracy. They forget that Gen AI isn't an automation tool; it's an augmentation tool. At the other end of the scale, RPA cannot handle ambiguity and couldn't use probabilistic thinking to generate content because it's built for structured, repeatable workflows. It's also worth noting that any form of AI might lack situational awareness or environmental context. 

It is crucial to define the task and how the technology supports the human performing it. Any investment in AI should be made against a backdrop of a solid understanding of foundational emerging concepts like these. 

We need to do some (more profound) thinking

Beyond blind investment, there's also a need to think strategically about the difference between how a company approaches AI adoption at an enterprise level and how its legal department should integrate AI into its operations. This requires careful consideration of AI's capabilities and how it can be safely and ethically embedded into processes, systems, and procedures. 

Solving the conundrum

An excellent way for businesses to answer "Could this work? And is it worthwhile investing in?" when it comes to AI is to ask the teams who will use the tools to weigh in early and often.  We recently chaired an 'AI Shark Tank' for a client who is determined to be an early adopter. Their legal and risk professionals pitched ideas around how AI could transform business processes. One idea was using a Gen AI chatbot to triage legal department queries. This simple but highly effective AI application may save them hundreds of hours yearly.

Risk shouldn't be a trade-off for speed

For legal practitioners and in-house teams, AI adoption is never just about enhancing work products, creating space for critical thinking or generating efficiency - it's also about managing risk, governance and compliance. Confidentiality, privacy regulations, cyber risks, IP and regulatory obligations require rigorous oversight. Being in breach should never be the trade-off for speed.  Anything that goes wrong with AI is inherently systemic. This means one point of failure will permeate the organisation, presenting a new level of risk to regulators.

Emergent value 

At Ashurst, several early practical AI use cases emerge, including high-volume work (detecting patterns and flagging inconsistencies), contract life-cycle management, litigation analytics, enhanced e-discovery use cases and regulatory monitoring. Our Risk Advisory and Ashurst Advance team are advising a growing number of clients on how to safely and effectively apply emerging capabilities to achieve a strategic advantage without the associated risk of compromising compliance. 

Whether you are already embedding AI strategically or still in the planning phase, the key distinctions between 'using AI' and 'embedding AI strategically are both understanding its capability and pursuing clarity on its scope. Those who get this right will lead the future of legal innovation.