Your Artificially Intelligent Agent

The moment that Watson became the Jeopardy champion will be remembered as a watershed moment in human history. It is the Gutenburg Bible of our time.

Media interest in Watson seems to have reawakened this month, with IBM now talking about how exactly Watson is being implemented to help make difficult medical diagnoses. In this area where humans just can’t process the amount of data necessary to make the best decision possible, Watson could be truly revolutionary. For now, Watson’s work will still need to be checked by meatspace doctors to make sure he doesn’t send some poor soul off to have a limb amputated unnecessarily, but this does not take away from the awesome power of data processing this will be putting into the hands of doctors.

This got me thinking about how life is going to change when we all have access to our own Watson. This raises important questions about how much control over our lives these intelligent agents will have. If you could have a hyper-connected computer intelligence which could monitor everything you say and do to check for errors and help you to consistently come to the best decisions, wouldn’t you want it? It is not only the high level decisions of doctors that could benefit from more data integration, but every decision we make as individuals could really benefit from a little more intelligence.

Lets take a simple example – the run of the mill job search in 2025


Instead of writing a resume, you task your artificially intelligent agent (AIA) to investigate job opportunities. Under your direction, AIA will explore the internet for available job opportunities which might mesh well with your specific skills and knowledge. The stored knowledge base of your AIA might also play into the decision, as you two are really a team at this point and that should be taken into account. Other factors your AIA might take into account, are your interests or willingness to move, where your family and friends live, how much time you have available, how much money you want to make etc…

The beauty of getting your AIA to do all of this work is that it can make a decision which balances all of the many factors involved with a finesse that a lone human would never hope to achieve. So in just under an hour, AIA brings up 7 draft emails to companies she proposes she proposes would be a good fit for you.

You see that one job is in Sydney, Australia. Your ex-wife lives in Sydney. Do you really want to open that can of worms again? Does AIA have an eye on your romantic life with that option? Maybe AIA just wanted you to consider whether you are interested in rekindling some connection there… maybe if you decide against this it will lower AIAs ranking of putting you in a situation possibility of reconnecting you with youjr ex-wife… You have been down this road before, there is no point spending time trying to figure out what AIA is thinking… Scratch Sydney.

So of the remaining 6 emails: 4 are applying to open job postings that might be interesting to you, 1 is an email reaching out to an old friend at a company you used to work for to see if they are looking for anyone, and the last email is to a company that doesn’t have an open posting but that AIA inferred from public financial data that it could improve its profitability by 112% by hiring you. In short order, you look over all the emails with the attached optimized resumes and send them off.

The companies all have much more powerful AIA helping their decisions, so within a day or two you have heard back from all of those places you applied. Two of the applications were turned down flat, including the hail mary AIA sent to the company with no official posting. AIA is aware of the emotional impact that such a rejection might have on you, so she softens the blow by pointing out that there is no way to know what crazy strategy their corporate AIA has come up with.

AIA prioritizes what she thinks is the best offer and shows it to you. Its a short contract in San Francisco, the pay is good, and they only need 3 hours a day (under the normal 4-5 expected in many offices these days). AIA also shows you the better salary and work agreement she thinks she can negotiate for you.

“People used to do all this themselves, what a headache that must have been” you think to yourself, shaking your head in nostalgic disbelief. You go ahead and tell AIA that you’ll accept the job in San Francisco.


Having an intelligence like Watson in your pocket will change every facet of your life; from something as mundane as submitting tax forms or figuring out what to have for dinner, to the highest level social and medical decisions. You cannot overestimate how significant of a change it will be to have a powerful artificial intelligence agent looking over your interests. Ultimately, the power of decision will still rest in your hands, but an AIA lets you better see the map of where those decisions lead. 

In the end, with the emergence of AIAs we will address something we hate to admit to ourselves. We are shitty decision makers.  We do our best to weigh the relevant factors and come up with what we think will be the best plan, but we can really only take into account a fraction of what we should. How many ills of the world come about from bad decision making compounded upon mistakes? Some people might fear the loss of control associated with having an AIA help in your decision making, but at the end of the day the AIAs will make us smarter, and a smarter world is a better world.


40 thoughts on “Your Artificially Intelligent Agent

  1. This is a wonderful story, which I enjoyed reading, but you’re bending the limits of what is currently possible (even with rooms full of computers) with AI. I know it doesn’t seem like much of a stretch to go from “can diagnose illnesses” to “can assess resumes”, but it’s a vastly different problem. It all comes down to the number of variables that need to be tracked, and also something about the medical industry you might not know: they’ve been gathering and quantifying the data that Watson needs for decades. Watson can safely infer information, based on scads and scads of records, such as “these 600 patients who had blood in their phlegm with this illness reacted poorly to certain painkillers” because those records are digitized, the big step forward Watson has taken is being able to extract information from loosely defined medical records and turn them into processable, actionable data. It did the same thing with wikipedia and other various online sources of information in order to win Jeopardy. The key is that Watson is unable to come up with new information, it can’t think, it can just quickly process information that other people have already made available.

    Most of the things you have included in your story are feasible, and will probably come to light given time, which is exciting. Some of them (profitability assessment, automatically assessing candidates) we still need some work on.

    Keep the thoughts coming, it’s the one place human’s will always shine.

    • ” it can’t think, it can just quickly process information that other people have already made available.” And what exactly do you think we as humans do? Base all of our decisions on what constellation the sun is rising in this year? We make our decisions based on past experiences (not saying this is a good way but it is what threes bay majority off people do). Job hunting, especially during a downturn in the economy v is a huge pain and takes s severe amount of time. If I have a “search” running fort me, i would be thrilled. I simply can’t justify spending 12hrs a day/5 days a week doing nothing but job searching because that is what it takes now. Google alerts only go so far and the sheer number of job sites to search is insane.

      Regardless , what concerns me about the medical side is that the knowledge is severely limited to “western” practices, which have been shown time and time again to treat symptoms and never the root cause. Companies forcing untested drugs by bribing doctors, reactive decisions instead of teaching people at a young age.

      The old computer addage stands: “Garbage in, garbage out.” Slanted knowledge is slanted.

      So really, Watson is just a really expensive Wikipedia then?

    • *@” The key is that Watson is unable to come up with new information, it can’t think, it can just quickly process information that other people have already made available…”

      >And that part inofitself I think is the scary part , for me anyways, to think something(not even someone..) that can’t think; would have my life(literally) in their hands(well not real hands lol!) if this comes to exist. Though doctors also base their opinion on books, etc…least they can think quickly to make quick decisions if need be. The invention of technology is great when used properly, for some things, but there are still instances(as in this case) I’d prefer a real thinking human! But the post and thought process was thought thanks for sharing. 2 thumbs UP

    • What does it mean to make a decision? What does it mean to think?

      These are key questions which come up when we think about AI. Is the compter really thinking? The thought experiment of the Chinese Room, is a good analogy of the problem we are looking at here.

      • It’s a really interesting connundrum. As living, breathing humans, I think we’re biased in favor of those inputs into decision-making that seem most exclusively… human: emotion and chance. Do these un-computer-like inputs actually lead to better outcomes?

        To bring it back to the realm of the job search, my emotional hangups about rejection and ridiculous aversion (bording on phobia) or making phone calls gets in the way of me seeking better opportunities. I’d LOVE a computer agent to boldly represent me out there in the marketplace.

  2. On my blog, I have written a great deal on consciousness, AI, and “optimal decisionmaking.” I have come to find that what you deem as “smarter,” really is just us becoming the machine. And sure, we will run longer and more efficiently if we let ourselves become the machine, but will it be worth it? We will live longer, but will we be living?

    Very thought provoking, which is always my cup of tea.

    Julien Haller

    • I have sometimes entertained the idea that perhaps this is the time when we are living as the most liberated and powerful individuals the world has ever and will ever see. In the past we were weak victims of a natural world we did not understand and could not control, and in the future the boundaries that allow us to be individuals will fall away and we will become part of the machine. As you put it, “We will live longer, but will we be living?”

      But then again, are we really individuals at all now really. No man is an island. Maybe we as individuals are just the neurons of the wider world, with ideas firing between us at the speed of sound, or light?

  3. Will the AIA limit or increase options? The scenario makes me feel that possibly although the number of options may increase there is the risk that these might narrow our vision and reduce our ability to consider total changes in direction.

    • A good question. Will the margins of acceptability become softened or hardened? Will we be more or less free? A good question for a follow up blog posts I think.

  4. interesting, im intimidated by the prospect to be honest–regarding the diagnosis capability of “watson”
    on AI i have always envisioned this prospect….however i cannot help but think that AI is itself nothing more than man controlling machine through programming? or is it possible that at some point the machine may control the man? either way i think the guys creating this technology need to consider maybe using this technology for worthy causes- like eletric power or something of that nature or at sustaining our natural resources.
    we cant eat or drink computers, but computers or artificial intellengence may help in developing a better programme directed towards natural resources and conservation thereof.
    Im not a tree hugger and dont believe in this “global warming” drivel, im assuming a practical stance. We have enough technology to make our lives easy, lets creating technology to enable longetivity of our species

    • I think in the current climate of recessionary gloom and doom, people have forgotten that we can do anything we want. We can make the world however we (collectively) decide. The world is full of people who look at how things are and they say “This is what is possible, and the rest is impossible”. By my thinking, impossibilities come from lack of creativity and vision.

      If we want development to be done in a sustainable way, we can do it. All that we need is the vision of how to create systems that can make our dreams happen.

  5. As of late I have had a growing curiosity. Where is our technology taking us? So many variables, So many possibilities.Your story illustrates an ideal world for humanity.. And in one sense of the word improving our quality of life. But for me these changes come with some reservation.

    In my time I have witnessed the advent of the computer.The rational any machine is to do all the menial tasks freeing man from the mundane. Allowing him more leisure thus improving his quality of life. Clearly machines have done just that. we have come to depend on machines more and more. This does not come without problems. A small illustrate my point. Go to the grocery and the cashier rings up your items. But without the machine they are unable to figure simple calculation in their head head. That troubles me. Maybe I am just old school and consider learning as important.. The brain need it’s exercise to grow. All I am seeing is lazy brain. It is like we are cultivating a generation of mindless human.

    I realize the change is already upon us and it can be very good. we need to rethink how we are to use this technology. Will we use it or will it use us?

    • I am not sure it is so much “Lazy brain” as “Adapting brain”. The requirements of the old world are no longer relevant in today’s society. Just as knowledge of the best place to hunt rabbits is no longer determinant of your success in the modern world, an ability to remember highly accurate and specific facts is falling away. Instead, we are cultivating the ability to think abstractly about the world.

      What happens when computers can think abstractly though? This, I don’t know, but I would imagine it is going to lead us into a strange world indeed.

      • I did struggle with the lazy brain words. Not finding another I just typed that in. The adapting brain seems to be relevant . I guess it is just hard for me to let some of these things go. Like I love my books. I have a Kobo reader and the computer of course. But there is just something about having a book in hand. The way it looks, feels, smells. The whole concept of turning pages. Maybe I am the last of the bleeding heart romantics. It’s just me. Time will continue to pass with or without me.

  6. Very interesting article, while I agree with some of what you say I must disagree with your conclusion that we are shitty decision makers and that this technology will make better decisions. I offer a personal story as my reasoning: My grandfather worked for a company that made and sold bicycles for his entire career. His main job was to make the sales and revenue projections for the coming quarter. He did this for over 30 years without the use of a computer, even after the computer was almost a staple in every home and office. I remember watching him work once as he sat at his desk reading the newspaper, checking weather reports, and reading various different publications that seemed to have nothing to do with bicycle sales. He spent about 3-4 hours doing this and finally he looked up at me and say ok I got it! I never understood what he was doing until I was older but after he retired the company hired two young, hot shot IT guys to fill the position. They claimed that the algorithm they had developed could do my grandfathers job more accurately, faster and at half the cost. Till this day, and its been some years now, the company has never been able to come closer than about 20% to the actual sales and revenues but my grandfather would consistently make projections that came within 2-3% without ever using more than a basic calculator because he could make connections and factor in variables that the algorithm could not. He still loves to tell this story every time he reads an article praising the wonders and abilities of new technologies. The human brain will always be superior to AI in ways that are very much important but often overlooked or taken for granted, because there are some abilities of the brain that AI will never be able to have such as intuition, instinct or draw on past experiences.

    • A good anecdote about the current state of some algorithms versus human ability, but I have to disagree that this is going to hold true forever. If that algorithm really was sufficiently advanced, it would learn from its own mistakes and improve over time. It would also continually look for more sources of data, number of comments about biking on facebook, number of bicycle accidents in a given market etc… Eventually it would be looking at more data then a human could every account for, and I am confident it would significantly outperform a human. Every year computers get better, humans stay about the same.

      Really if you believe in math, then you believe that computers will eventually be better at doing things then us.

      • Very true, you make a really excellent point! While I don’t disagree with you here I would just add that an algorithm or any AI, no matter how sophisticated, is still only as good as the brain that designed and programmed it. My main fear is that, as we become more and more dependent on technology to do our thinking for us, the pool of creative minds capable of conceiving and designing newer and better technologies will grow smaller and smaller.

      • I don’t necessarily agree that the AI can only be as good as the brain that designs it. Depending on the task the AI can be many many times better for sure. For instance what’s the square root of 19 trillion? A computer can tell you.

      • I didn’t mean to imply that the AI couldn’t perform its intended function better than the designer could, obviously nobody could calculate the square root of 19 trillion faster than an AI. I just meant that if the AI isn’t programmed to incorporate certain variables in a calculation then the results it computes wont be as precise as they could be even though it can make the calculation faster than the person who designed it. There may very well be an AI that exists today that is capable of recognizing deficiencies in its own programming and add new search criteria or new variables by itself to correct the deficiency but I have yet to hear that this is even possible.

      • Maybe its not be possible yet, maybe it is, but given that you just conceived of it now you know for sure someone is working on it.

  7. I am reminded of the WWW trilogy by Robert J. Sawyer. It, like most great scifi is incredibly intuitive about what type of an effect an artificial intelligence would have if given access to all of the data on the internet. The story focuses around the relationship between a young girl and this all knowing entity that exists in the dead spaces of the internet; the girl names him Webmind and his growth from a snippet of code to a thinking entity is lightning fast! Great post! I will be following you:)
    In the least creepy way of course!

    • I agree that the WWW trilogy is an example of great sci-fi writing! I really liked the idea that separating two different connected networks can somehow stimulate the formation of a conscious mind.

  8. Pingback: Freshly Riffed 23: Dick Jokes Off The Port Bow! | A VERY STRANGE PLACE

  9. While some humans base their decisions on what constellation the sun is rising in this year (horoscope, anyone?), I’d like to think of the AIA as an optional device. Very optional. In answer to your question early in this piece, I absolutely would not want a hyper-connected computer intelligence to monitor everything I say and do. Even if it’s to check for errors and help come to the best decisions, I’d rather commit to my own failures than lose all ownership of my successes.

    And to be honest, some of the greatest things have come from stupid ideas. The city of Alexandria only came into existence because Ptolemy I stole Alexander the Great’s corpse out from under the noses of his bickering generals. Smart idea? Morally right idea? No. But one of the greatest cities in human history was founded because of it. Lastly, if a computer decided everything for us… Well, all of our decisions would be the same. They’d come into conflict with one another. Some of them could be outright terrible.

    There are approximately seven billion people on Earth, for example. If an AIA decides your best option is to end your life because there are two many people taking up resources– Do you do that? If, in 2025, an AIA is considered a solid authority, would you even think twice about doing it? I admit, this is an extreme example. But while I found parts of this article amusing, I found other parts of it disturbing. Mainly the very idea of a someone looking back in nostalgic disbelief on those crazy days when people would think and make decisions for themselves.

    …In an ideal world, this Watson / AIA plan might work on an individual level. As cool as Waston is, I don’t think it could function long for personal use. Then again, I bet people said the same thing about televisions and computers. And look where we are, right?

    • Would we really have any less capacity for decision making? AIA would lay out options and ultimately the decision would remain your own. In some ways you could say its better then what we have right now, wherein plain old luck decides your options from which you must choose.

  10. Ah, but to err is human. And we surrender something of our innate uniqueness when we opt out of decision-making. Oh, i do accept your premise that we are often appalling decision-makers. But are we any better at making intelligent products? The great danger of AI is, and always will be, that we humans (so frail and prone to error) actually created it/them. Beware what we wish for – it may come true!

  11. Very well said and well deserved in Freshly pressed. You have a very insightful post and I totally agree with your perspective. Having an Artificially Intelligent Agent has far reaching effect in our human life. We can get everything done efficiently and makes us smarter enough using advance concepts.

    • Another interesting thought while you are thinking about it, how much will the impact by multiplied when we can connect AIA directly to our brains.

  12. Nice post but there’s one question I would like to through out here. How would this affect the privacy of humans. Someone can use this to there advantage to control people. Nothing would be private anymore. It could happen in a lot of ways but two I think of are either hackers or the creators themselves using this to their advantage. Yes we advance but is the machine making the decisions or just the man behind it making you think we are so advanced but really controlled.

  13. Pingback: Your Artificially Intelligent Agent | Patrick Schaal's Blog

  14. Pingback: The Quantum Court: Assigning Blame in the Techno-future | Thought Infection

Leave a Reply to David M. Russell Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s