Make no mistake, there is a data science revolution underway. While artificial intelligence (“AI”) may bring currently unimagined benefits, unforeseen consequences will also manifest as a result of this transformation which is occurring at a pace beyond our ability to monitor, harness and govern. Organizations may realize exponential growth and new efficiencies and retail consumers may be dazzled and drawn to the benefits and new opportunities; however, they must also be aware of the potential pitfalls and exercise-informed decision making that is necessary to implement successful AI initiatives. 

The current reality is, most organizations have just begun to explore AI, so the time is ripe to inform and educate ourselves about the subject and the potential risks. The significance of the risks and the need for caution is becoming a concern for legislative and regulatory bodies. For example the July 18, 2019 Smart Brief’s ISDA Daily Lead online newsletter stated:

“The UK Financial Conduct Authority (“FCA”) is asking firms deploying AI and machine learning to make sure they have a firm understanding of what could go wrong with the technology. ‘Some firms haven't done any thinking around these questions at all -- which is obviously a concern,’ says FCA executive director of Strategy and Competition Christopher Woolard.”

In light of the optimism of those who rush to AI, and a growing recognition of the risks by others, four interrelated topics come to light:

• Algorithmic biases

• Human oversight and privacy

• Development overzealousness?

• Smart cities

Algorithmic biases

Biases in programming and machine learning can either form spontaneously or develop over time. They are commonly caused by development and coding errors, resource deterioration, stress due to heavy use and misdeed or a criminal act. These biases could eventually have a direct impact on the privacy, liberty, safety and security of individuals or groups of people. Once realized, these impacts could be difficult, if not impossible, to slow or reverse.

As algorithms begin to modify or write themselves in response to their increasing comprehension of real-word situations, the probability of unintended consequences increases. Even today’s most pedestrian AI algorithms are capable of making decisions many times faster than a human and they are already able to change the course of their decision- making, based on large volumes of real-time data.

Human Oversight and Privacy

This topic has always been a concern of the community; however, interest and emphasis of the importance of human oversight and privacy have significantly increased since Amazon’s recent admission that Alexa recordings are shared with QA ‘testers’ who then share ‘interesting’ recordings with other employees, among other bad behavior. Additionally, a recent news story about a murder that was captured live and the ensuing debate over its use as evidence has also helped to raise the focus on the issues.

Currently, three U.S. states have some form of consumer privacy bills that have been passed into law, and only ten more states have a bill in front of their governing bodies. None of these laws or bills have any provision specifically written for the kind of data that potentially can be collected by an AI-managed device.

Development Overzealousness?

It’s as simple as the age-old story of ‘just because you can doesn’t mean you should.’ Are we building at a pace where protections from development mistakes, spontaneous or curated, are unsustainable? Should there be regulations governing adaptability, extensibility and other mechanisms through which AI can learn and adapt? These are just a few of the critically important and currently unanswered questions in the public conversation regarding how much control we should relinquish to a computer system, or the computer developers and their corporate employers.

With the yet unimagined and potentially significant benefits to our society, for example aiding in curing diseases and eradicating global hunger, where do we draw the line?  Should it be with food production or with our criminal justice system?  Even if society is able to define a red-line, how does it get enforced and by whom?

Smart Cities

The concept of “smart cities” has been around for a long time. In a large number of cities around the world, immense quantities of data are captured daily via security and traffic cameras and other facial recognition and recording technology. The capture and use of this data has improved the safety of citizens and significantly aided in both crime avoidance and resolution in ways that simply didn’t exist prior to the technology implementation. At the same time, the use of the data, by whom it is analyzed and for what purpose are issues that raise important legal, privacy and consent concerns. In many cases, these are previously unchartered territories of the law and are currently being studied and analyzed by the prominent legal scholars across the world. What will eventually be resolved via the legal systems versus the legislative and regulatory systems continues to develop and evolve.  Regardless of whether it be by litigation or legislation, there is undoubtedly much more on the horizon when it comes to types of bias in AI and their impact on smart cities.

AI and the Law

How are law firms preparing themselves to better serve their clients as the adoption of AI becomes common place?

A number of law firms are looking both internally and externally at ways to apply AI.  Internally, they are deploying AI and other data science techniques to improve efficiency, effectiveness and profitability. An example is the trend to add Chief Revenue or Chief Billing Officers and supporting staff to their organizations over the past several years.  Some of these resources are market-facing and interact regularly with their firm’s clients. These resources are analyzing troves of billing and market data to provide their firms with a competitive and strategic advantage for pricing services and developing alternative billing solutions while concurrently improving profitability.

A less prevalent but also growing trend is the addition of non-traditional roles of Chief Data Scientist / Chief Data Analytics / Chief Data Officer and supporting resources at law firms.  These roles are looked at to provide a strategic and competitive advantage for the delivery of legal services to their respective firms’ clients. One of many examples is the application of machine learning tools and techniques to improve the efficiency, effectiveness and timely delivery of regulatory and investigative services.    

The legal industry revolves around precedent and case law. What do we do during times when no case law exists’ or new case law is developing? As recently reported by WSJ Pro’s John Murawski, law firms are actively preparing for AI legal cases by bulking up their ranks with artificial intelligence experts to secure new business. As reported by Murawski, issues include “rights to datasets, ownership of algorithms, liability for property damage and personal injury.  It’s only a matter of time before we see an increase in AI related litigation and for all of us in the industry, now is the time to educate ourselves and our clients to prepare and plan for the inevitable.”

In summary

The AI discussion will continue for many generations to come, with the tone and topics of the conversation changing at an exponential pace. In the not too distant future, there will be well-defined and established AI practice groups delivering services across the legal ecosystem.

Governing bodies will unquestionably move toward regulations, albeit far behind the development curve. If history is an accurate indicator, delayed responses will seek to regulate technology already abandoned for the next discovery and offer little real protections. Such attempts to manage the revolution will be swept aside as technologies outdate the language of the law.

What can we expect?

Companies that don’t even exist yet will vie for the top spot with technologies intended to ‘improve our lives’ and reduce our effort on mundane tasks. Consumers will continue the race to replace devices with the newest model, while not necessarily understanding the invasion they invite into their homes.

AI will find its way into every corner of our lives from this point forward. Future historians will analyze and evaluate how society proceeds in its adoption of AI in the coming years.  

We should expect to be impressed by the tools soon coming to improve our lives, give us access to unimaginable knowledge and to maybe one day cure diseases. These technologies will result in observable changes to our personal and business lives as we know them today. This area of technology advancement is of significant importance to all of us and now is the time for us to increase our awareness and further educate ourselves about both the opportunities and risks related to the adoption of AI.