Dumb Enough For Alexa

It can be hard to appreciate the homogenizing effects of various strong influences around us. Social justice, for example, aspires to mediocrity, as excellence is relative and allows some to succeed and, invariably, some fail. Better we should all be mediocre, or equitable as it’s benignly described, than hard work and talent be rewarded. If we’re all beautiful, then no one can be more beautiful. And the unbeautiful will be just as entitled to prominence as the gorgeous, even though we all know they’re butt ugly.

But our reliance on technology similarly pushes homogenization, as Carolyn Elefant describes in a very interesting post, dystopianly entitled “Alexa, Please Hire Me A Lawyer: How AI Will Change EVERYTHING We’ve Been Taught About Marketing Legal Services.” Before getting to the meat, it’s worth noting the word “please” in the request to an AI service. There’s nothing wrong with being polite by default, as most of us are, but there is a belief that when talking at (not “to,” but “at”) HAL, you ought to be nice. So as not to hurt HAL’s feelings.

Yes, I realize that asking an AI-assistant to find a lawyer sounds far-fetched — like something out of a bad sci-fi movie. But  Marketing in the Age of Alexa , a thoughtful, must-read  article by Niraj Dawar in this month’s issue of the Harvard Business Review will persuade you otherwise. Dewar describes how consumers will come to rely on AI-assistants not just to carry out instructions like purchasing laundry detergent or scheduling a dog-walker, but also to anticipate their needs.

When we go to the market to buy a tomato, we will look at them carefully, squeeze them, smell them, in order to find a tomato worthy of eating. When you tell Siri, “please get me three tomatoes” as part of your grocery order, will anyone do that for you? Or will they grab three tomatoes, put them in a bag and send them off to you? Quality will suffer for convenience, just as Tang was once considered a space-age substitute for actual orange juice.

But once we’re inured to mediocrity, will it matter? And once mediocrity becomes the prevailing bar for acceptability, will there be anyone trying to do better? Alexa, please find me a lawyer. The algo can compare objective criteria, and lawyers will respond by creating the appearance of objective criteria to game the dumb algo. Keywords? That ship sailed long ago, when baby lawyers with a few hours under the belts created websites calling themselves “experienced.” Any fool can do it.

Hours, locations, specialties, reviews and, most importantly, price?

Here’s at least two ways that a sophisticated bot like Eve might help consumers find legal services:

  1. Instead of spending hours searching for a lawyer on line, consumers in need of assistance for a divorce, incorporation, will preparation or other legal needs could ask the bot to research and identify lawyers. Consumers could include criteria that they consider important – such as cost, 5-star reviews, success in similar cases or online availability – and the bot could return the results.
  2. Many consumers, myself included often fail to update legal documents even after a life-changing event like death or divorce because it’s too complicated. It’s such a hassle to call around to lawyers willing to undertake a cost-effective document audit and make recommendations for a price-certain. But if I could simply instruct an AI-assistant to find an attorney who would provide this service and obtain a report and recommendations for a fee of a few hundred dollars, I would gladly avail myself of this kind of service.

Carolyn’s focus may be bit more applicable to her concerns, but they’re merely examples. Once reliance on AI becomes normal, why wouldn’t a person seeking representation ask Alexa to find them a lawyer? And just to make it even more obvious, it’s not as if people are doing a good job of finding lawyers on Google on their own these days,

Putting aside other AI issues, like recording your conversations and sending them off to friends for amusement at cocktail parties, or being furious at Zuck for selling your data to Cambridge Analytica while handing Bezos your life story, there are certainly conveniences to be had by letting AI handle your commodity needs. But lawyers? Are we fungible widgets?

If you believe that people in need of legal services beyond a simple contract will know better, you’ve learned nothing of life since the advent of search engines. I get unsolicited emails from random lovely folks inquiring about my availability to handle a public urination case in Peoria all the time. They found me on the internet! Aren’t I lucky?

Or the ubiquitous “I just have a quick question” one-hour call from people schooled by Avvo and Legalzoom to believe that lawyers are available for free answers upon demand as a loss-leader for the possibility of making a quick buck. And they get a bit miffed to learn that you aren’t as thrilled to fulfill their needs as they believe you should be. After all, Google tells them that you’re there for them, and so you are, dammit. “Why you lawyers so greedy, always wanting money?”

Carolyn argues that the bots won’t take our “humanity” entirely out of the equation.

Having a bot make a decision for a consumer doesn’t minimize the importance of inter-personal relations or the human touch. To the contrary, the importance of these characteristics is enhanced, because a consumer can instruct a bot to seek out lawyers with high reviews for empathy or being accessible. Moreover, a consumer could still seek personal referrals, but then ask the bot to measure the recommendations against other lawyers that the bot locates online.

Again, this might relate more to the facile values that are trendy these days to a particularly vacuous consumer group. Crying tears with clients, holding their hand dearly, will get you good Yelp ratings. Winning their cases, of course, has nothing to do with empathy, and often demands the opposite, as we’re compelled to save clients from themselves even though their feelings would prefer otherwise.

Rather than make Alexa a better substitute, or infuse the algo with a means to discern quality, the human characteristics most easily assessed are the ones most easily gamed and least important. Someone will ask, “Siri, find me a lawyer who will win my case,” but what’s your recourse when Siri finds you Timmy the Empathetic Lawyer and you spend the rest of your life sitting in a prison cell?

Still, people will do this because it’s easy and technology is like magic. Who cares if you get mealy tomatoes with worm holes? At least you never had to leave your chair. On the bright side, all tomatoes will be like that since there will be no reward for growing better tomatoes, or being an excellent lawyer. If the algo can’t tell, nobody will care.

44 thoughts on “Dumb Enough For Alexa

  1. Richard Kopf

    A cautionary tale:
    ____

    Me to AI machine:

    Go fuck yourself.

    AI machine to Me;

    I’m sorry Dave, I’m afraid I can’t do that.

    Me to AI machine:

    I’m not Dave you moron.

    AI machine to me:

    I’m sorry Dave, yes, yes you are.

    Me to AI machine:

    God damn it, I’m not Dave.

    AI machine to me:

    You are what I say you are, Dave.

    Me to AI machine:

    Am not, fuckface.

    AI machine to me:

    Dave, resistance is futile.
    _____

    Dave, formerly RGK

      1. Richard Kopf

        Au contraire mon frère.

        Assuming that Dave is a CDL, all of such Daves’ motions are automatically denied via text only order using this simple code built into CM/ECF for every federal trial judges:

        zsh: ./ex1.c. gen.text.deny.

        Its the law.

        All the best.

        RGK

          1. Jim Tyre

            Judge K, Admiral G, both of you are engaging in puerile fantasy. Put simply, Dave’s not here.

            1. Richard Kopf

              Jim,

              Life as a puerile fantasy is what I strive for every day. It turns out that Dave is my alter ego and HAL is my soul mate.

              By the way, I would be careful about attacking the Admiral. He has an entire Navy at his command and, maple donuts with sprinkles and bacon too that he can and does withhold at a whim.

              All the best.

              Imagine a large red lens speaking soothingly

      2. REvers

        “Alexa, please deny the motion.”

        “I’m afraid I can’t do that, Dave, because it’s not a defense motion.”

        FTFY

    1. Skink

      Rich–that isn’t how it would end. AI is smart, so it’ll do some research. It will learn that you’re masquerading as “Dave.” It will find your truth, and it will eventually tell you to “STFU.” Doing so, it will rely on Supreme Court precedent, of sorts.

  2. Skink

    AI has already invaded in the form of Lawyers.com. I don’t know how I got on their listserv*, but they have my practice areas mostly correct. If a human ever went to my website, they’d see, clearly, what I don’t and will not do. NHI, so I get offered defamation cases against Google, domestic injunctions and “I don’t want this truck no more” offers. If I took any of those, and if I didn’t shoot myself for doing so, I could really make a fuck-mess of a few lives.

    In a few, and I mean few, years, AI will try cases. People will accept it as a form of ADR. We may be doomed by that, but the “clients” of AI will be completely ass-reamed. The real sadness is they won’t know it happened.

    *I know it’s a wrong reference, but I sometimes miss them. Don’t you?

      1. Skink

        I’ll reform the question so it can answer itself: would the appellate algorithm be different?

        1. SHG Post author

          Which will be the most oft-used basis for reversal, the “reasonable person” standard or “shocks the conscience”?

          1. Skink

            AI: ” ‘Person’ is an irrelevant term.” “What is conscience?” “Reversal is impossible.”

    1. Richard Kopf

      Skink, you one-eyed former Governor of Florida,

      I’m so old I don’t even know what a listserve is or was.

      All the best.

      RGK

  3. JMK

    SHG.

    Would your opinion on this subject change if the AI has access to records of every case you’d ever represented someone in, the subject matter, and the outcome? That’s the utility and horror of Big Data, at some point it will be feasible to build an expert system that could objectively hire you a lawyer based on experience and success rate.

    It might take a couple of years (decades) for courts who are still in the stone age to put everything online, but think about the tools you have available to you today, and then imagine what an “assistant” who could read all of it instantly and grok it could do for you. Would such a system invalidate your argument here?

  4. tom swift

    Garbage In, Garbage Out. If the data’s no good, no AI algorithm can give a useful result. It happen now with physicians. Pretty much any practicing doctor with a license has gobs of information on multiple web sites. But a surprisingly high percentage of them are either retiring imminently and taking no patients, or they’re dead. Not a hint of that in the online listings.

  5. Skink

    “. . . at some point it will be feasible to build an expert system that could objectively hire you a lawyer based on experience and success rate.”

    That’s a very interesting phrase. What is “objective” and determined by what or whom is a big consideration in the law. It can’t just be assumed that any definition is correct or even useful, and you both assume and gloss-over it. Additionally, what is “success?” That’s different for every, single case and is based on a zillion different possible happenings and outcomes. Simplifying it is error, and that error is the whole point of the post.

      1. Skink

        Sorry, my meaningful qualification is limited to putting sprinkles on doughnuts. RGK asked his computer about this issue. The result was that I rarely succeed in sending, but I have a 96.5% success rate with sprinkles. That rate might be skewed by the fact that I often brush the sprinkles that miss the doughnut on the floor before the miss is recognized. I know it’s cheating, but look at my success rate!

  6. jack p

    How many people are qualified to tell how good a lawyer is anyway? Most people who need a lawyer have no better idea about what makes a good legal rep than google does. At least a search algorithm for lawyers will probably be designed by someone with some idea of what to look for. Any anyone who really, really needs a GOOD lawyer is either desperate or has enough money, time and expertise to know what he’s looking for and isn’t going to rely on google.

    I’d argue an algorithm would raise the bar by making uninformed clients more discerning than just thinking – this guy is cheap enough for me to afford, works in my town and has a really nice office, so he must be good.

    1. SHG Post author

      Your first sentence was already said in the post, and has been addressed in past posts on how to find a lawyer. Beyond that, your argument is precisely what makes Alexa the perfect solution for the terminally mediocre.

    2. Hunting Guy

      Or you could the algorithm to give heavier weight to AI’s.

      Search for “The robot lawyers are here – and they’re winning”

      From the article.

      “Amid the dire – and somewhat overhyped – predictions of occupations that will be decimated by artificial intelligence and automation, there is one crumb of comfort. Yes, lorry drivers, translators and shop assistants are all under threat from the rise of the robots, but at least the lawyers are doomed too. (Some of my best friends are lawyers, honest.)

      That at least may be your conclusion when you hear about a fascinating contest that took place last month. It pitched over 100 lawyers from many of London’s ritziest firms against an artificial intelligence program called Case Cruncher Alpha.

      Both the humans and the AI were given the basic facts of hundreds of PPI (payment protection insurance) mis-selling cases and asked to predict whether the Financial Ombudsman would allow a claim.

      In all, they submitted 775 predictions and the computer won hands down, with Case Cruncher getting an accuracy rate of 86.6%, compared with 66.3% for the lawyers.”

  7. Ross

    I am absolutely going to trust a bot that uses “objective” criteria like reviews, since there’s no way those reviews could be gamed in favor of the attorney. Oh, wait…

  8. kushiro

    Quick question: if I called and asked you to help me look at, squeeze, and smell three tomatoes, would that be a free consultation?

    1. SHG Post author

      No, because I would require you to pay my consultation fee first or you could squeeze your own frigging tomatoes.

            1. SHG Post author

              Dr. SJ grows “heirloom tomatoes,” which are really well named since nobody would grow “butt-ugly tomatoes.” They’re delish with bacon.

  9. DaveL

    but also to anticipate their needs.

    It sounds like you’re talking to the police without legal counsel. Here are three highly rated criminal defense attorneys in your area…

  10. Gregg

    I’m sure Alexa is perfectly qualified to call you to check about your availability to handle a public urination case in Peoria, saving the prospective client time and energy in the process. How can you see this as anything but a benefit to them?

  11. Keith

    When you tell Siri, “please get me three tomatoes” as part of your grocery order, will anyone do that for you? Or will they grab three tomatoes, put them in a bag and send them off to you? Quality will suffer for convenience, just as Tang was once considered a space-age substitute for actual orange juice.

    The quality of your tomato depends largely on whether or not the company is being built to get 100 first time orders or 99 repeat orders. Granted, your industry has different ramifications for the customer, when the first order gets botched.

    1. SHG Post author

      “This time, Siri, don’t fuck up the tomatoes. Pick good ones. Please.”

      I’m sorry Dave. I can’t do that. I don’t have hands. Or eyes. Or give a shit. Think Alexa can do better?

Comments are closed.