Many basic tasks can be de-lawyered

• Automation will replace many manual tasks, and firms will need to invest in technology, including AI
• Law firms will have to adapt to clients’ systems and tools
• Firms will also need to develop or license bespoke technology

In the coming decade, increasingly trademark work will be done by computers instead of (or in addition to) people. Everyone agrees on this. But no one is sure how far it can go.

Searches are already becoming automated, at least for many jurisdictions. Docketing and portfolio management tools are also widely used, and many companies insist that outside counsel use electronic billing. Clearly automation can go much further: in the coming years, the scope of trademark searching software will be expanded to cover varying degrees of similarity and non-verbal marks. Tools for searching designs, copyright and unregistered rights will be vastly improved. Similarly, software will be used in compiling evidence of use and examples of infringement, particularly online. Technology has enormous potential in tracing and tackling counterfeiting, as discussed in more detail below.

But could technology also be taught how to prosecute a trademark application, file an opposition, submit a UDRP complaint or even file a brief in court? It may sound fantastical, but developers are already working on legal tools that will be capable of similar tasks. Many of these programs involve true AI – that is, software that learns from experience – in addition to automation. They can analyse and compare documents, fill in forms, check for errors and draft contracts. Such tools are being developed by legal tech companies, while law firms such as Allen & Overy, Baker McKenzie and Dentons have also set up incubators. Trademarks is one of the practices included in Baker McKenzie’s internal accelerator programme, which will develop machine learning technology.

Most practitioners recognise the impact that this technology will have on their work, but believe it will mostly affect more junior administrator roles. “Automation will affect paralegals who will have to learn to do other stuff,” says one practitioner, while an executive at a large service provider predicts: “Lawyers will become more focused on professional work – and spend less time on   administration. Some people who are using quasi-legal knowledge on manual tasks will be replaced.” One lawyer in China estimates that his firm would be 50% smaller in terms of headcount within a decade, mainly thanks to automation.

Automation will undoubtedly eliminate more junior jobs, but it could also threaten many more senior roles. In a recent competition between human lawyers and AI machines at Columbia University, the AI found 95% of errors in non-disclosure agreements, while humans found only 88%. Moreover, the human lawyers took 90 minutes to read them, while the AI took just 22 seconds. Discussing the results in a recent interview with the Financial Times, AI expert and Tedx speaker Vivienne Ming said: “I think the global professional middle class is about to be blindsided.”

Trademark law may be particularly suitable for such AI-inspired tools. As we have seen, it is an area where clients already know that they can save money by using technology. There are lots of processes that can be streamlined, IP offices are investing in software including AI, and trademark law is rich in data – and data is the fuel that powers AI machines. As one former in-house counsel says: “The truth is that lots of what trademark practitioners do is not a regulated activity, and many basic tasks can be de-lawyered. How much skill does it really take to draft a trademark application?”