Since the beginning of the internet, the potential for technology to be exploited and misused has been a primary concern for policy makers. As technology has continued to expand and play an ever more prominent role in our everyday lives, these concerns have only broadened, becoming both more frequent and unique – and with each new technology comes a host of new ethical concerns.
As technology continues to evolve in the business world, while the legal profession may have been slow to adapt and progress, times are changing. The development of technologies such as machine learning and artificial intelligence are being applied to legal in an effort to reduce overheads, improve accuracy and boost efficiency. But despite the advantages legal tech brings, there remain concerns around technology and how it may implicate professional ethics.
In the research that underpinned this report, 18% of those surveyed said that implementing new technologies within their teams had raised ethical questions. On its surface, this would suggest that in the legal space, the use of technology was almost always done in an ethical fashion and didn’t raise concerns. But speaking in more detail with some of those counsel who had faced ethical issues painted a different picture, suggesting instead that many general counsel are not yet attuned to the new types of ethical questions arising.
Ethical by design
Just how technology is developed for, and used within, various legal functions is one source of significant ethical debate – particularly given the inherently secretive nature of the profession. One general counsel working in the banking sector, describes some of the ethical issues encountered with how legal software is applied.
‘In Brazil, we have a major problem dealing with the volume of mass litigation cases involving consumer and labour issues that we face. Here at the bank, we currently have over 120,000 active lawsuits.’
‘To help deal with this, I have software that allows me to access cases to find out what judges are doing, or have done, in cases related to my company and/or my competitors.’
Having software that enables counsel to quickly view other similar cases from the past is not revolutionary, but having cases analysed in order to provide suggested arguments according to the jurisdiction and the judge is.
‘The information is public, but this system can offer me suggestions on which versions of arguments would be best for certain judges. It can also help with predicting what may happen in my case.’
‘Using this, I can build a better defence or consider alternative options. For example, if the software revealed that I have a 70% chance of losing a case, I can use this information to propose that we may be better off agreeing to settle the case. Then, when it comes to working out a settlement, I have the statistical information that would allow me to know how many other similar cases have been settled. That provides me with the information that will help me decide how much money I should settle a lawsuit for.’
Using statistics and predictive capabilities in themselves is not inherently unethical, but the potential to build on this system in unethical ways is apparent. The potential for alternative sources of data that could be used to help build a better profile on judges and predict their behaviour is vast – social media is just one prominent example of this.
‘This kind of concern is being raised, but it is not a major concern right now. What may be an ethical concern is the data itself. The real problem arises only if you use confidential information, or if the data could be used to make correlations,’ says Marcelo de Araujo associate professor of philosophy of law at the Federal University of Rio de Janeiro.
De Araujo has worked closely with developing ethical protocols and codes for emerging technologies, including machine learning. He says that the data used to teach these systems can itself be an unintended source of ethical issues.
‘Machines can be influenced by human bias and emotions. If the original data is biased, the output delivered by the machine will also be biased. Machine learning will typically find out underlying patterns in the data set, a task that would take too much time and expertise for human beings to do unaided by AI systems,’ he says.
‘But if the original data set contains an unusual number of references to decisions that, for instance, marginalise black citizens, chances are that the prediction suggested by the AI system will reflect that kind of pattern.’
Alejandro Fernández R-B, head of legal at Cotemar, has grappled with similar issues.
‘In Mexico, when you have a labour case, mainly it’s because the former employee feels that he was entitled to receive a bigger compensation. Some cases, they are right; some cases, they are wrong.’
‘But the software might recommend that I can settle with them for a lower figure that they are entitled to. So, that is the ethical issue. Because in some part, I want to, and it’s my duty with the company, and on the other hand of course, I want to be fair with the former employees. So, that is an ethical issue that I am seeing as a result of this software.’
Show me the data
Since the implementation of the General Data Protection Regulation (GDPR), safeguarding data has not only been an ethical responsibility for general counsel, but a legal obligation. Many countries across Latin America have followed the lead of the European Union and aligned their existing regulations or introduced new laws around data protection.
Brazil’s suite of data protection laws are set to come into force in August 2020. The Lei Geral de Proteção de Dados (LGPD) stands to have a marked impact on business and broader society, with protections possibly set to be enshrined in constitutional law in the future.
‘The public at large increasingly perceives the protection of their personal data as a pressing question. There is now a new bill for an amendment to the Brazilian Constitution that would turn the protection of personal data into a fundamental right,’ explains de Araujo.
‘Most citizens seem to be aware that they are giving away too much personal information online, and that it is far from clear what private companies are currently doing with their personal data. Over the last few months, there have been reports of two cases of technical failures which compromised the personal data of 28,000 citizens in one case, and 70 million in another case.’
Fair and Available
There are a number of ethical rules that apply to general counsel, irrespective of technology. However, as legal tech becomes more prevalent, new ethical dilemmas begin to arise. In Brazil, AI tech is being introduced to reduce the time it takes to settle disputes.
‘The Brazilian supreme court announced in 2018 that an AI system called Victor was being tested in order to speed up court decisions,’ says de Araujo.
‘Currently, the court may take as many as ten years to settle a dispute. Many citizens die before they can obtain the benefits of a favourable decision.’
Victor is an artificial intelligence tool developed in partnership with the University of Brasília. The AI software is designed to process thousands of court decisions already made and identify links between cases. By identifying and analysing previous precedents, the aim of Victor is to increase the speed of legal proceedings through the development of neural networks. Technology has the potential to play a major role in improving legal services in Brazil, but de Araujo believes that algorithmic bias needs to be accounted for.
‘There has not been much public discussion on how Victor is supposed to deal with algorithmic bias. Speeding up court decisions would be a huge improvement in the Brazilian legal system. But quicker decisions will not promote justice for ordinary citizens unless unjust patterns of decision-making in the past are also avoided in the future. For now, it is unclear how the Supreme Court intends to address this issue.’
To avoid transferring prejudicial precedent to future cases, more oversight and transparency within machine learning services is required.
Says de Araujo: ‘The problem could be addressed by examining what exactly the system is doing as it provides scores for risk assessment. But developers may be unwilling to make their AI systems more transparent. They fear that, by making the system more transparent, other developers might take advantage and copy their technology.’
In both of these cases, the culprit was a government institution, heightening public awareness of data-related issues at a time when business is preparing for increased scrutiny with the introduction of the LGPD.
‘As a banking service, we provide credit cards to a lot of our clients. Using this data, I could learn, for example, that a client of the bank is going to the beach every weekend, because they buy something on the way there,’ says the banking GC.
‘The specifics of what you buy is information that is protected by law, but a bank could use other information from the transaction to try and sell the client a product. For instance, you could contact the client to try to sell them travel insurance, because you know that this person goes to the beach every weekend.’
Equally, it could be said it would be unethical if a bank were to use this information to discriminate against a client by developing assumptions based on spending patterns.
‘People who go to the beach every weekend could be discriminated against if I make a correlation with something else that affects their credit score. The client could call the bank and ask, “Why is my credit score affected?” Right now, if I say that our system says you go to the beach every weekend and your salary is $1,000, therefore you probably cannot pay your bills, we would face problems,’ explains a GC in the financial sector.
‘If we make a lot of correlations like this, we can face problems. This kind of model can become so sophisticated – it deals with a lot of data and one of the dangers is that this data could trigger a warning that says this person is a good or bad client for the bank. I think we could face real problems in the future.’
In order to address the potential for making such ethical and legal breaches, banks across Latin America have implemented measures to limit data access across a company.
‘We have a specific department here at the bank comprised of lawyers, software engineers, operations people and risk analysts. They analyse and see the flow of all data in the company and check to see if the information is sensitive or if it must be kept confidential,’ he says.
‘This team applies the law in order to protect the information of the people linked to these types of risks. I think we have this department in the bank after the implementation of the GDPR law and we are spreading this culture of data protection in banking – we are not only compliant to banking secrecy law, but for data protection. There is still a lot to do, but the bank is in a good shape’.
In the future, the implementation of new technology will only continue to make data collection and analysis more sophisticated and complex. Data protection has always given rise to ethical concerns for corporate counsel, but technology has made data security a major point of concern. With data breaches gaining global attention, some general counsel in Latin America believe that it will only be a matter of time before stricter global controls are implemented.
‘I really believe these types of regulations will become global at some point. Maybe the United Nations or another global entity will try to implement regulations that are more globally focused, but I would guess that, regardless, the principles surrounding data protection will all be the same,’ says Pablo Enrique Urrego Hernández, head of legal at Diageo Colombia.
Will Siri practice law?
Separate to the ethical concerns raised from algorithmic bias and data protection are the fears surrounding legal tech’s potential to replace lawyers.
‘Some people believe that there is an ethical discussion surrounding legal software that could attempt to destroy the value lawyers bring,’ explains Urrego Hernández.
‘If you develop software that could replace lawyers, then essentially a lawyer could be replaced by a machine. The idea of technology is not to be viewed as a risk to job opportunities, but instead as a tool that could help lawyers focus on tasks that are much bigger and add more value.’
Despite the way legal tech is viewed, its potential to be an industry disrupter is generally accepted. Of the more than 200 general counsel surveyed across Latin America for this report, 84% thought that technology would be a moderate to great disrupter over the next five years.
‘What will probably happen is that systems will be so highly developed and so detailed on what we want them to do, that tech will probably not take into account any human emotion. There are so many factors and possibilities that could be used to create a system that could probably emulate repetition, but, in the end the human factor is key,’ says Urrego Hernández.
‘At the moment, there is not true artificial intelligence – I do not know whether this could change in the future – but, up until now, there has been no system that can replace human behaviours and ways of thinking.’
Legal tech across Latin America, like much of the world, is still in its development stage. Most of its implementation has revolved around low-level work but, as technology continues to develop, the issues that could be raised by legal tech are still fundamentally unknown.
‘Experience has taught me that while you develop these things, you start seeing issues you have never seen before, and there is no doubt that the future of legal tech will present many barriers that will need to be addressed. At the moment, what we need is to be objective, although imagination is powerful and out-of-the-box thinking is good,’ says Urrego Hernández.
‘The list of ethical concerns that could be raised by new technology is extremely extensive. What should happen is that ethics should be part of your life. At the end of the day, ethics should be one of those main considerations that you have to take into account in every aspect of your life.’
Whether in-house counsel across Latin America resist or embrace technology, legal tech will inevitably transform the delivery of legal services, though not without a host of other issues to contend with. Ethical conduct will remain pivotal to ensuring the legal profession remains independent, effective and accountable.