Karanovic & Partners
  April 19, 2018 - Serbia

The Use of AI in Law: How Just is Machine Justice?

Artificial intelligence is a burning topic in many sectors today and the legal industry is no exception. Recently, at the World Services Group’s annual employment law1 conference held in February, AI was heavily debated along with its’ impact not only on the legal profession, lawyers, clients, the way business is done, but also our traditional understanding of concepts such as “law” or “justice”.

According to expert predictions, with current software developments and the rise of intelligent systems, in a few decades certain jobs will disappear – bank clerks, cash-register operators etc. It may be difficult to imagine, but it is actually not that far from reality.

Artificial intelligence and employment

Software is even now used by HR professionals in the recruitment process and in later stages – certain interviews are conducted with no humans present apart from the interviewee, with intelligent software asking the questions. Whether by the various backdoors or through the main gates, technology is replacing the human workforce. So, what are the consequences?

Certain jobs that seem necessary now will become obsolete – what will happen to the people doing those jobs? If they would become redundant, the system would need to find them new jobs or enable them to re-qualify in a timely manner. However, there are many positive aspects to this – such as the development of new professions and industries. In a sense, even though some jobs will disappear, new ones will pop into existence to replace them. Of course that there will always be jobs that machines are not sophisticated enough to perform, that will always need a “human touch”. On the other hand, when working in dangerous environments, machines clearly have the advantage.

Further, from an ethical viewpoint, the subject of discrimination in the relations between humans and their machine interviewers is quite interesting, especially in the recruitment process. Can we get accidentally discriminatory solutions based on the input “fed” to the programme on the acceptable candidate? Artificial intelligence still eventually requires humans to operate it and feed it information, so can a machine then really be more objective than a human?

Other legal repercussions of the development of AI are quite substantial. If, for some reason or other, the machine fails to perform adequately, who is then to blame? Traditionally the one held responsible would be the user or the owner. Now, perhaps the one to blame is the developer that worked on that AI?

Due diligence software

The legal profession is no exception to these developments. A growing number of international law firms use AI to conduct due diligences, partially or in toto.

For example, in the most limited use of software for due diligence, the programme is given information from a virtual data room, and through a combination of key-words it eliminates unnecessary or irrelevant documentation. The next phase in more developed DD software includes the programme processing the relevant documentation through the recognition of certain word combinations or information, pairing it then with its’ own legal terminology based on the laws and rulings that it was fed. The AI can also compare information, fill in parts of the report or assists in some other way.

Since most clients these days request a fast reaction and a due diligence more and more often needs to be conducted within few days, in order to speed things up and reduce costs, international law firms decide to use this this kind of software as it makes things a lot easier, since long and arduous work can be sometimes resolved in hours.

Online dispute resolution and AI judges

A topic which excites considerable discussion is the development of online dispute resolution. In cases when there is no need for hearings, and when disputes are handled through submissions, the current development of technology enables that submissions by the parties can be simply uploaded and read – after which a ruling can be reached, without the need for the physical presence of the parties before the authorities. Most courts are still not as advanced as to use this method, but there is a tendency to implement online submissions in some more flexible forms of arbitration or mediation. Technically it is a very small step to go from filing your submissions in hard copy to uploading them.

This brings us to perhaps the most controversial application of AI in the legal industry – software passing judgements and rulings. The majority of lawyers would say that this is the last area that they would expect to be conducted via software. Most would say that this requires justice – a characteristically human concept.

However, it may come as a surprise that a similar kind of software is being currently tested by some judges and professionals in EU as we speak. Most notably, this option is currently being considered for collective lawsuits, where numerous claims arise from the same factual background and the judges are actually copy-pasting their decisions in practice.AI can do this faster and more efficiently, saving the judges’ time for more complex decision making. This appears to mark only the beginning of the silent AI revolution in the legal segment, since we may not be that far away from AI making decisions in other typical cases as well.

An artificial intelligence judge has accurately predicted most verdicts of the European Court of Human Rights. Scientists built AI that was able to look at legal evidence and asses its ethical questions to decide how a case should be decided. And it predicted those with 79 per cent accuracy, according to its creators. The algorithm looked at data sets made up 584 cases relating to torture and degrading treatment, fair trials and privacy. The computer was able to look through that information and make its own decision – which lined up with those made by Europe’s most senior judges in almost every case.

Paradigm shift

Would such a use of AI make law more neutral and unbiased? Is justice dealt out by a machine is more objective than human justice? Of course, this relates to the question of whether we want law to be objective in such a way or do we want to look at the specific circumstances of each unique case from a human perspective? Then again, some would argue against AI objectivity by pointing out that humans feed it information so that it can never reach the ultimately objective approach.

The idea is that we still need to feed the software with court rulings, precedent law and information from which it would develop syllogisms, reach decisions and propose further actions. So, while inputting information, certain discriminatory ideas could be included accidentally, allowing for discriminatory rulings. According to some researchers, for example, republican judges are more inclined towards stricter punishments than their more liberal counterparts. Also, according to simple statistics, judges reach harsher verdicts before lunch than after. And, as much as judges try to be objective, there are indications that courts are generally more forgiving towards women or more inclined to provide them with certain rights – especially mothers. So, despite our attempts at neutrality, certain patterns emerge, given a large enough sample, that the AI can pick up and continue with in its decision making process.

So, will machine justice be more just than human justice? And if so, would we want such a thing?

The AI judge in our example was wrong in around 20 per cent of rulings – its logic and conclusions were mostly correct, similar to those of its human counterparts. Taking that into consideration and given the cost-effectiveness of the use of software as opposed to using human lawyers – the final word could be left to the clients. Whether to hire lawyers or to opt for the more affordable solution, even though not ideally just.

Conclusion

Until recently, the legal profession was much the same as it was two thousand years ago. From the times of Ancient Rome, we have had the prosecution, the defence and arbiters that consider the arguments and then pass judgements.

We are witnessing now a paradigm shift – the ideas of law and justice are being reconsidered. How we adapt to emerging technologies, new client expectations and industry trends will determine the place of lawyers in this brave, new world.

1 employment law