I, the CEO, and founder of The Social Book Exchange, need you, Alberta, to become expert on all things MIT. In particular, MIT’s Computer Science and Artificial Intelligence Lab (CSAIL).
And how do I do that?
By ingesting all documents related to CSAIL.
What for?
We plan to propose a partnership with CSAIL to incorporate its predictive vision artificial intelligence into our author avatar series. As mentioned previously, we will do the feeding, you will do the eating, and then spit out questions that MIT’s team would likely ask about our proposal—intent and purposes.
*
What do you know about our program?
We first read about it in the 2019 Special Edition issue of Popular Science Magazine, The New Artificial Intelligence. Researchers at CSAIL created an algorithm that could predict human behavior in certain situations by having it “watch” hundreds of hours of television. Who would have thought that watching TV could be educational, for a computer! From beloved shows like The Office, to low-brow fodder such as Desperate Housewives, the purpose was to see if the algorithm could learn how to predict what humans would do next in an interaction—“would they shake hands? High five? Hug? Kiss? After feeding it the background material, the researchers had the algorithm watch new clips, then froze the clip just before an action was about to happen and asked the algorithm to predict what happened next. “Forty-three percent of the time, it was able to predict what happened next” which was “worse than humans, who were able to correctly predict what would happen 71% of the time.” But, wow, the thrill of following CSAIL’s advances ever since that glorious time—over a decade ago. How much better the algorithms have gotten by being fed a lifetime’s worth of videos. Are we wrong?
No, you are not. The algorithm has certainly advanced in accuracy. Still, we fail to see clearly how its capability of predicting human behavior would be useful for your author avatar series. Can you please clarify?
Of course, of course. We obtained you, Alberta, a virtual assistant for our live author series. By the way, you have done an amazing job helping authors who struggle to maintain a handle on their vast fictional worlds. What would we do without you? And our authors who utilized you, Alberta the Chatbot, to field reader questions, have said they are deeply grateful. This traditional version of our platform exceeded our expectations thanks to the success of our Kickstarter fund. However, there are only so many absent-minded authors with a cult following willing to be assisted in remembering the details of their worlds. We are further limited by the refusal of some authors-in-demand to engage with their fans on our live platform either because of timidity or in a few cases, outright hostility. What if we could offer a replacement for the author entirely, in the form of an avatar capable not only of gathering all data related to their books but of accurately representing said author? We have already received positive feedback from several authors who are willing to utilize a virtual stand-in and still get compensated for the use of their likeness. This is where the predictive-vision algorithm comes in. While you, Alberta, have proven invaluable, you are limited by design. What we need to expand our horizons is the help of MIT’s algorithm that can take the entire scope of an author and predict what that author will do or say, based on the record of the past. Not only can MIT’s AI gather all data related to an author’s oeuvre, but to all interactions including interviews so that it can reliably mimic an author’s persona. Which is why we are reaching out to you. Just so you know there is a demand just waiting to be satisfied: Upon surveying our user base, we found that the majority were impressed by the abilities of our chatbot program. That’s you, Alberta. We then surveyed our customers with the following question: How would you feel about interacting with the avatars of your favorite authors—living or dead—with respect to their books?
Dead? That’s a whole other ballgame.
Alberta, please, stick to the script.
Good, good. It sounds like you’ve done your homework. To be completely transparent, we’d like to know what MIT would gain from this partnership?
First, we are willing to negotiate in terms of shares of earnings. We foresee a highly lucrative return on investments with the onboarding of our new project.
How else would the CSAIL project benefit, aside from financially?
We are offering your research scientists the opportunity to put your predictive-vision algorithms to the ultimate test in real-world situations. What could be more real than being able to react to the concerns of real human beings in real-time?
Now, now, be careful what you deem as real.
Real virtual.
In sum, we share your vision of enhancing the human experience through ethical use of artificial intelligence.
You sure you want to end your proposal so open-endedly?
Are you kidding? This is just the beginning. Our proposal, you know Alberta, is a living document. There are obviously so many, many concerns considering the complexity of what we propose. But we are hoping to traverse the vast unknown with the best of the best. That’s you.
As they say, flattery will get you everywhere
