I’ve been crushing on Alexa for months, but didn’t get up the nerve to buy an Echo Dot (the model that looks like a hockey puck) until Amazon rolled out a call feature where you can call from Dot to Dot. (If you’re not familiar with Alexa, she’s a personal assistant, like Siri on your iPhone, but on a speaker in your house.) So now I have four set up in our townhouse, and I don’t have to yell anymore to get my teen and tween daughters to come downstairs for dinner. “Alexa, call Julia,” “Alexa, call Maya.” Maya, my youngest, asks Alexa to tell a joke, “Alexa tell me a joke.” Alexa happily complies, “Which baseball team would win at tug o war?… The New York Yankees.”
Now, I’ll be sitting between client calls and feel in the mood for Chopin. “Alexa, play Chopin,” and his “Nocturne In E Flat Major” begins to ring out. It’s fantastic. She also can do more useful things, like give me a summary of the news in the morning or recite my calendar items to keep me up to date. It’s easy to think, “How in the world did I ever listen to music before?” Hmm, I actually had to open up an app, type in a search for Chopin, and press play. I harken back to when I was a kids in the 80s and the remote control came along, except this is voice control… for almost anything.
Alexa and Siri are examples of personal assistants or chatbots. They are cool and useful, but until recently, they were the only game in town and built as closed systems, where only Apple and Amazon engineers had the tools to create dialogs and functionality.
What is a Chatbot and is It Artificially Intelligent?
A chatbot is a computer program designed to mimic human conversation. In 1950, Alan Turing posited that if a person, communicating by text alone, could not tell if the other participant were a computer or person within five minutes, the computer could be deemed “intelligent.” This is the Turing test.
That all changed in April 2016 when Facebook opened up its Messenger (messaging) platform to chatbot developers. Unlike Alexa and Siri, the bots available on Facebook Messenger are (for now) entirely text-based. Shortly thereafter came text-based bots on Skype, weChat, Telegram, Slack, Kik, Line, and SMS. If you’re not familiar with some of these channels, you’re not the only one, but trust me, a lot of people use them. At Facebook’s annual F8 conference this year, Mark Zuckerberg announced that developers had created 100,000 chatbots. Many people think of bots as the new apps. They don’t require a download and they work on whatever operating system you use, be it Windows, Mac, or Android. Now, Amazon has made it possible to add “skills” to Alexa, so she can do everything from set a timer for billable hours to recite daily affirmations. Apple is working on releasing its own AI-speaker product (like Echo) with a dedicated chip to improve its artificial intelligence.
You might say, “Okay, enough already, I’ve heard the hype about AI and chatbots. What does this have to do with me?” Let me tell you. Last year (beginning of 2016), I was tooling around on Google News to stay current as one does, and I came across an article about this kid in the UK who had built a chatbot that asked a user a series of questions and created parking ticket appeals. That kid is Joshua Browder, now a Stanford student, and his service is DoNotPay.co.uk. Remarkably, this computer code had been successful 64% of the time, defeating 160,000 parking fines and saving drivers millions of dollars. I remember feeling goosebumps and thinking to myself, “This is going to change everything.”
Now, you may think me a little melodramatic. I mean, how many lawyers make a career out of parking tickets? Parking ticket defense is high-volume, low-margin work that most lawyers avoid because it would be too much hassle to make a living that way. And, that’s the genius of it. It’s a story we’ve heard over and over again. A disruptive technology targets a low-value product or service to gain a foothold in an industry, then leverages that success to challenge incumbents for higher-value clients. Apple disrupts IBM with personal computers, Toyota disrupts GM with economy compact cars, Canon disrupts Xerox with inexpensive personal copiers, and the examples go on and on. Yet, we are now living through this disruption in the legal industry and making the same mistakes as prior incumbent businesses.
Would You Rather Talk to a Chatbot or a Human? Survey Says…
Nearly 70% of consumers would rather ask a chatbot a question over a human because they prefer to receive an instantaneous answer.
Do people really want to be talking to disembodied chatbots? Don’t they value the personal touch of a handshake and speaking face-to-face? According to a recent study of UK consumers commissioned by Ubisend, when asked what’s most important when communicating with a company, 68% say it is ‘reaching the desired outcome’, closely followed by ‘ease of experience’ (48%) and ‘speed’ (44%). Being able to contact a brand at the time most convenient to them is also important to 39% of UK consumers. When asked “Why would you consider asking a chatbot before a human,” nearly 70% said they would prefer to receive an instantaneous answer. As we might expect, consumers care about results.
At least three pillars define the scope of legal work: legal documents, legal advice and court advocacy. Chatbots promise to automate the first two. It is this difference in kind, not degree, from prior disruptors, such as LegalZoom and Rocket Lawyer, that makes chatbots a unique opportunity and danger to the legal profession. Already, we read innovation success stories, such as Visabot, which offers tourist visa extensions, artist visas, H1b visas and help with DACA, works kind of like TurboTax and has processed over 50,000 visa applications since the end of 2016. Yes, over 50,000 applications in less than six months.
Certainly, chatbots are early in development and raise practical and ethical concerns. Common objections to chatbots are that they are not yet truly “intelligent,” do not pass the Turing test, are artificial intelligence based on good old-fashioned rules-based systems (not machine learning), have been done before by expert systems in the 70s and 80s, and are generally not perfect. (I call this last one the panacea objection because it has as its premise that if a technology does not provide a perfect solution, it is not worthwhile.)
The response to all of these objections is that they are all correct. The current incarnation of artificial intelligence employed in chatbots is rules-based, it would not pass the Turing test, and it is not perfect. But these objections miss the point. Chatbots do not have to be intelligent (in the Turing test sense of the word) to be effective. They can be programmed once with legal domain expertise to provide a service, and can thereafter scale to assist hundreds or millions of legal consumers by means of a smartphone.
Legal objections to chatbots raise the risks of waiver of attorney-client privilege, the unlicensed practice of law, conflict of interest, and more. These are all legitimate objections that need to be addressed, but none of them are show stoppers. For example, confidentiality issues can be avoided by defining the service as a non-attorney service. But, I believe we lawyers should be careful not to get trapped in microanalysis and lose the forest for the trees. I understand it’s a natural reaction for lawyers; we’re trained to issue spot. But we cannot and should not let the perfect become the enemy of the good.
A few use cases, or practical examples, of how you might incorporate chatbots (and their conversational interface) into your legal practice to make your office run more efficiently are as follows:
- Qualifying potential new clients
- Onboarding new clients
- Identifying and triaging legal issues
- Scheduling appointments with clients and others
- Conducting client interviews for fact gathering
Lawyers Who Remain Relevant Will Succeed
[T]hose lawyers who recognize the pending changes as an opportunity will likely do very well in this new environment. ―Roland Vogl, executive director of CodeX in The Coming of Age of Legal Technology
Chatbots are a proxy for automation generally, and how we react to chatbots will dictate how lawyers remain relevant in light of technological change. So what can be done in the face of this paradigm shift? How can lawyers practically respond? My advice to you is to ignore the doomsayers and the specter of robot lawyers, and be the change you want to see in the world. Embrace disruptive innovation. Experiment with legal technology and see how you can make it work for you. Learn more about what’s going on and get involved.
Here are some concrete ways to do that:
- Follow CodeX. CodeX is the Stanford Center for Legal Informatics and it has been focused in this area of legal technology for over 10 years. Its LegalTech Index stores a curated database of over 700 legal tech companies to explore. And, if you sign up for CodeX’s mailing list, you will get invitations to its weekly meetings with legal tech founders who present their latest innovations.
- Go to a Legal Hackers meeting. Legal Hackers is a group dedicated to developing creative technological solutions to legal problems with chapters worldwide. Legal Hacker groups usually consist of lawyers, programmers, students and policy makers, and meeting formats range from coffee socials to workshops to panel discussions. You can find your local chapter by doing a search for “legal hackers” on Meetup. If you don’t find a chapter in your city, you can always apply to create a new chapter.
- Attend a legal tech conference. Luckily, this isn’t as much of a challenge as it used to be. CodeX has its FutureLaw Conference, there’s the ABA Tech Show, Lawyerist’s TBD Law, the Clio Cloud Conference, ALM’s LegalWeek, Lawyernomics by AVVO, and EvolveLaw has many meetings throughout the year. Whichever conference you choose, make sure you meet somebody new and learn something you had never heard of before.
- Follow #legaltech on Twitter. If you’re not active on Twitter, don’t worry, just follow the “legaltech” hashtag and you’ll learn about the latest innovations, tools and strategies in legal technology. You’ll also start to get a handle on who cares about this stuff and follow them as well so you can see what they post on the subject. Twitter is also a great way to stay in touch with those new friends you make at the conferences.
- Get your hands dirty. Create a chatbot. And remember, it’s okay if it’s not perfect. A free and simple service you can try out is ChatFuel. You can literally create a chatbot that’ll work on Facebook Messenger in 10 minutes, watch this video to learn how. Later, you can try out more sophisticated chatbot creation platforms, like Motion.ai, API.ai, LUIS, Wit.ai, and many others that seem to be launching every month.
Upon completing this article, I asked Alexa to turn off Chopin and whispered to her, “Alexa, can I hire you to be my legal assistant?” If you meet me sometime, ask me and I’ll tell you what she said.
About the Author
Tom Martin is a legal tech advocate, author and speaker. He is the founder and CEO of LawDroid, an AI and chatbot consultancy and development company, and is co-founder of Vancouver Legal Hackers. Follow him on Twitter @lawdroid1.