LexisNexis: AI has changed legal work. It hasn’t changed who is responsible for it.
Artificial intelligence is reshaping legal work at a remarkable pace. Tasks that once relied entirely on manual research, drafting and review are increasingly supported by AI, prompting organisations to rethink established workflows, introduce new ways of working and reconsider the skills their people need. Much of the conversation around legal AI naturally focuses on what the technology can do and how quickly those capabilities are advancing.
Equally important, however, is what AI has not changed.
The responsibility for legal work still rests with the people producing it. Professional judgement, critical thinking and accountability remain fundamental to legal practice, regardless of how much technology supports the process. As AI becomes more deeply embedded in everyday work, those principles are becoming more visible rather than less.
That theme surfaced repeatedly during a recent discussion between Richard Tromans, Founder of Artificial Lawyer, Oliver Bethell, Chief Technology Officer at Travers Smith, Nigel Lang, Chief Information Officer at Fieldfisher, and Karen Waldron, Director of Product Development at LexisNexis. Although each panellist approached the subject from a different perspective, they consistently returned to the same conclusion: AI is changing how legal work is carried out, but responsibility for that work remains firmly with people.
Governance, implementation and organisational change all featured during the discussion, yet the strongest thread running through it was how organisations preserve judgement, transparency and accountability as AI becomes embedded in legal work.
Watch ‘Take back control: owning AI risk, governance and adoption’ to hear the full conversation
Standing behind the work
The panel quickly moved beyond the capabilities of AI and towards the responsibilities that surround its use. While AI is undoubtedly changing the way legal work is produced, none of the panellists suggested it changes the professional obligations attached to that work. Instead, they repeatedly returned to the role of human judgement in reviewing, challenging and standing behind AI-assisted outputs.
Oliver expressed this in straightforward terms when discussing the use of AI within Travers Smith: “The software doesn’t change the fundamental principle that our lawyers are responsible for the product that they provide.”
Nigel reinforced the same principle from a governance perspective: “The human needs to be in control of the AI, ultimately, and accountable for that.”
Although they approached the subject from different perspectives, both arrived at the same conclusion. Confidence in AI depends just as much on the people using it as the technology itself. Professional judgement remains the point at which legal expertise, context and accountability come together, and those qualities cannot be delegated to technology.
Our recent In CTO we trust research reinforces the same theme. Lawyers are far more likely to use AI confidently when they can verify outputs, understand how answers have been reached, and remain assured that responsibility for legal work continues to rest with the individual exercising professional judgement.
Responsibility doesn’t end with the lawyer
While responsibility for legal work ultimately rests with the lawyer, the panel also recognised that people should not be expected to carry that responsibility in isolation.
Nigel challenged the idea that responsibility for AI naturally sits with technology teams, describing this as one of the biggest misconceptions organisations are making. He argued that supporting the responsible use of AI depends on expertise from across the organisation rather than a single function taking ownership.
Professional judgement is exercised by individuals, but it is influenced by the environment in which they work. Clear expectations, appropriate oversight and opportunities to challenge or verify AI-generated outputs all help reinforce the responsibilities that lawyers already hold. Those organisational factors do not replace professional accountability; they help people fulfil it consistently.
The panel stopped short of presenting a single model for how organisations should approach this, and rightly so. Every organisation will develop its own structures and processes. What remained consistent was the principle that responsibility may belong to individuals, but organisations have an important role in supporting them to exercise that responsibility well.
Designing for responsible AI
Responsibility featured throughout the discussion in another important way. Karen highlighted the contribution technology providers can make in supporting responsible AI use. If people remain accountable for AI-assisted work, the technology itself should help them exercise that responsibility with confidence.
Rather than relying solely on training or policy, Karen argued that responsible AI should be reflected in the way products are designed. Features such as clear citations, transparency around how answers have been generated and straightforward ways to verify outputs all make it easier for lawyers to review, challenge and stand behind their work. As she explained: “Vendors have a responsibility as well to design systems to make responsible use of AI the easy path.”
That perspective complements the wider discussion about professional accountability. Human judgement remains central, but good technology can strengthen the way that judgement is applied. When verification is built into the workflow rather than treated as an additional task, responsible behaviour becomes part of the normal way people work.
Confidence in AI is influenced by more than the quality of a model. It is shaped by the experience people have when using it, their ability to understand and verify outputs, and their confidence in taking responsibility for the decisions that follow.
Learn more about the RELX Responsible AI Principles
Judgement remains the differentiator
Towards the end of the session, Richard raised a question that brought together many of the themes explored throughout the discussion: how do organisations prevent people from becoming overly reliant on AI?
The responses were revealing. Rather than focusing on future advances in AI capability, the panel consistently returned to the factors that shape how people use the technology today. Education, transparency, verification and thoughtful implementation were all discussed as ways of helping professionals develop confidence in AI without diminishing their own judgement. Technology can support better decisions, but it cannot replace the experience, scepticism and contextual understanding that legal professionals bring to their work.
Every stage of the AI journey, from selecting appropriate tools to reviewing outputs and advising clients, ultimately comes back to people making informed decisions. AI may change how legal work is produced, but the responsibility for deciding whether that work is accurate, appropriate and fit for purpose continues to rest with the individual.
The panel made a compelling case that confidence in AI is built not simply by adopting new technology, but by creating an environment where professional judgement continues to be exercised, supported and valued.
Watch the on-demand webinar to hear the discussion in full
