updated on 12 December 2017
QuestionArtificial intelligence is on the rise, but how much of a threat does it pose to junior lawyers?
Artificial intelligence systems (AI) have undergone a relatively unnoticed transformation from science fiction to an integral part of our daily lives. Our habits and behaviours are monitored, recorded and analysed so that our digital personal assistants can be more efficient, shops can make bespoke recommendations and heating systems can warm our homes to the perfect temperature for when we walk through the front door without being asked. AI’s impact on business is indisputable, and this extends to professional service providers such as law firms.
Most transactional law firms operate via a tiered structure, with a small number of senior lawyers who concentrate on complex and deal-specific issues supported by a larger number of junior lawyers dealing with a high volume of analytical work. Given both its focus on process and speed when compared to human work, AI is in many ways suited to undertake tasks traditionally given to junior lawyers.
In certain sectors, contractual arrangements are becoming more standardised. Trade bodies such as the British Venture Capital Association and the Loan Market Association have precedent documents aimed at standardising contracts across their respective industries, which eases review and facilitates swift negotiation based on precedents. Legal AI has been developed which relies on the standardisation of documents and uses it to recognise and analyse commercial contracts, as well as suggest amendments based on a pre-determined position a law firm would advise a client to take. Clients seeking to maximise efficiency and control costs can select legal counsel who in some instances replace junior lawyers with well-trained legal AI, and particularly with those which will produce a simple and reader-friendly work product, rather than a lengthy mark-up.
For law firms with clients which are global businesses and therefore subject to numerous regulations governing a vast number of transactions, legal AI can be developed that can search sanctions databases or antitrust legislation and determine whether transactions should be approved or will have to be refused due to potential breaches of applicable regulation. Here, legal AI offers two principal advantages over junior lawyers – it is in many cases far quicker and more cost effective at data collection and analysis than a junior lawyer, and it does not suffer from any potential human error that may creep in.
However, there are some significant drawbacks from reliance on legal AI. Primarily, the quality of legal AI is entirely dependent on the legal inputs used in creating them. Clients will ultimately need to rely on the accuracy of legal AI, in some cases with potential liabilities running into the millions of pounds. If legal AI providers are unwilling to compensate their clients for losses based on reliance on their products, it is unlikely that they will be sufficiently trusted to be used for any high-value work. Given the challenger and start-up nature of many current legal AI providers, it is unlikely that they would have deep enough pockets for clients to recover any losses from them, even if providers were willing to compensate such loses.
This risk is reduced when using junior lawyers at a reputable law firm. Law firms will train their junior lawyers sufficiently well as to minimise the risk of errors, and will likely structure deal teams in such a way that junior lawyers are not taking any decisions which are beyond their experience levels. Equally, senior lawyers can, and in most cases will, review junior lawyer work product to minimise any reputational risk caused by junior lawyer mistakes or inexperience.
Second, a systematic approach to analysis will inherently sacrifice flexibility in order to provide a certain and consistent outcome. While there is a growing trend towards standardisation of documentation, legal AI will struggle with analysing both (i) more bespoke documents which deviate from market norms, and (ii) documents which are drafted in such a way that they will not be detected by legal AI, due to effectively using synonyms or language that falls outside the scope of the legal AI’s knowledge. Also, advice from legal AI will often not be tailored to specific scenarios due to the legal AI’s inability to account for similar but differing situations, such as varying levels of bargaining strength between counterparties, or a range of counterparties with different priorities (eg, founder managers and public companies).
Equally, legal AI can provide accuracy on a certain date only to the extent that the legal analysis underpinning the system has not changed since the date that the system was created. Subsequent legislation, recent case law or revised compliance guidance from regulators will therefore need to be worked into systems on a frequent basis. Any time delay between the effective date of such update and the update of the legal AI will render the legal AI to a certain extent obsolete and therefore not a product many clients would seek to rely on. In cases where a settled interpretation or application of such update has not been established in advance of the effective date (eg, the current uncertainty over how best to apply the MiFID II Rules when they come into force) the viability of the business of legal AI providers will be seriously disrupted.
Junior lawyers will be instructed on the various nuances and differences between different client types and fact patterns, and will therefore be able to differentiate between, and provide more appropriate advice on, a broader range of transactions than legal AI. On time delays, while law firms will also suffer, most international law firms have the resources to be in the position to release client updates for significant updates on the same day as such updates come into force. In many cases, senior lawyers are involved in the drafting of updated legislation, and therefore have the requisite knowledge and expertise to quickly opine on new updates and present them to clients succinctly and contemporaneously. Junior lawyers can therefore be briefed internally and brought up to speed faster than legal AI can be reconfigured by a third party legal AI provider.
While law firms traditionally adopt a conservative approach, there are numerous examples of leading law firms using, and in some cases developing, legal AI to operate alongside their existing junior associate base. A hybrid approach allows law firms to offer its clients a competitive pricing package due to the reduced time spent on traditionally high fee areas such as due diligence and contract review, as well as offering its junior associates involvement in more complex and business critical work streams.
Law firms working in partnership with legal AI providers can ultimately provide clients with the comfort they need around the quality of the legal AI, which both reduces the likelihood of any losses caused by reliance on it, as well as increasing the prospect of recovery in the event of any such losses. Rather than viewing legal AI as a threat, junior lawyers and law firms should seek to adapt legal AI to their advantage to ensure that they can continue to provide innovative and market-leading services to their clients.
Joe Bradley is an associate within the EMEA private equity team at White & Case.