AI: Regulatory uncertainty
Hesitation about adopting artificial intelligence (AI) by law firms nervous of the risks presented by a regulatory grey area is causing a bottleneck that needs to be cleared, according to LawtechUK.
Although the Solicitors Regulation Authority has advised it was unlikely to take regulatory action in cases of errors caused by machine learning (ML) as long as firms have done what they can to avoid them, it was unclear exactly what steps must be taken, the body said.
LawtechUK has published a consultation paper seeking to work out exactly what concerns regulated law firms had about using the technology that might be inhibiting them.
It said: “A perceived lack of clarity of certain rules is… constraining the deployment and development of ML by authorised firms and individuals.”
It identified two particular areas that needed clarifying: how law firms can meet their professional duty of confidentiality to train and build ML systems cost effectively and how professional indemnity insurance (PII) cover might be affected by the use of ML.
LawtechUK pointed out that, because regulation was outcomes-focused, it was up to a law firm to decide what level of assurance it should obtain to demonstrate it had done everything it could to prevent errors occurring.
Specific risks using ML included bias in the input data, client privacy and security issues, data quality and confidence in outcomes.
It concluded: “A greater understanding of how these risk factors materialise and their severity in the specific applications of ML in legal services, as well as the risks this presents to consumers, is needed to support responsible use of ML.”
Questions included what precautions were currently taken, whether PII considerations affected the decision to use ML, what techniques were currently used to ensure confidentiality and whether there were other issues of concern.
The report highlighted the many potential advantages of ML for legal businesses and how an extensive range of products was now available, including those produced in-house by law firms, such as Shoosmiths’ Cia, a contract review tool.
External vendors ranged from big providers like Thomson Reuters to small start-ups such as Lexverify’s ML assistant, aimed at helping companies eliminate legal and compliance risks in their electronic communications.
LawtechUK also emphasised how ML applications have become a normal part of wider daily life and the practice of law, such as Google’s ubiquitous search engine and its use in legal research and e-discovery.
It said: “AI, specifically [ML], is transforming how law firms and departments operate and businesses and people interact with legal services.
“From legal diligence and regulatory compliance to the advent of next-generation chatbots, machine learning brings many opportunities to improve outcomes for businesses and society.”
The consultation closes on 5 February.
