Artificial Intelligent Attorney: The Future of Legal Research?

JUNE 10, 2016

Could lawyers soon be replaced by robots? Take Ross, for example, a young in-house attorney at IBM. He spends most of his time reading through legal documents, presenting hypotheses when asked to, and putting together responses along with references and citations to back up his conclusions. Could his job be replaced by Artificial Intelligence? Well, yes and no – because Ross already is a robot. Based on IBM’s cognitive computer Watson, Ross is “the world’s first artificially intelligent attorney”.


Unlike existing technologies that simply search for keywords, Ross’s makers say he is able to mine facts and conclusions from over a billion legal text documents a second. It learns from experience, gaining speed and knowledge the more you interact with it. It can also work alongside you, monitoring the law for changes that can positively or negatively affect your case.[1]

So is Ross a threat to your future career, or a useful ally? Lara Vivas, lawyer at Cuatrecasas, Gonçalves Pereira’s Barcelona office, believes that “there are a number of activities that lawyers do that could be done more efficiently and faster by computers, such as the review of documents that can be time-consuming and of low value. Having ‘AI lawyers’ could help the rest of the firm to concentrate in added-value activities.”

AI could also be useful, says Vivas, “in jurisdiction comparison, where an ‘AI lawyer’ can be much more efficient and faster than a set of lawyers calling to review a certain contract or policy. You just click a button and they could tell you whether there are issues with a certain clause in a certain country or state.”

Vivas plays down the fear of robots substituting lawyers in real life, because “activities where cultural elements make a difference will be hard to code, such as negotiation. Also, it is difficult to substitute humans in strategic decision-making, where AI can be of very relevant help in terms of providing information, but not necessarily in making the decision.” However, she does add, ominously, “at least with the current AI models.”

The need for legal firms to offer added value has already become more important, says Vivas: “There will probably be more and more services that the client will be able to get online almost for free, so the legal profession will have to move to added-value services and be ready to invest in technology.”

In a similar way to the “smart contracts” that already exist, whereby contracts are auto-executed, AI will be a useful tool. However, a bigger danger might be over-reliance on technology. Vivas raises the prospect of discrimination. While robots may seem inherently objective, their codes have still been written by human beings.

“For example, in terms of recruitment, there is already quite a lot of technology helping to sort out resumes,” says Vivas. “Programmers decide which data to feed to the machine to help draw conclusions and design the algorithms... Therefore, the apparently objective result of the machine can actually embed a discrimination decision and perpetuate biased decision-making.” So if you do find yourself working alongside Ross, just remember that he is fallible too.

[1] Ross. The future of Legal Research.