Last week, The Globe and Mail reported on a "new" Rembrandt, a painting imagined by computer algorithms and painted by machine to resemble as much as possible an image the Dutch painter would have created during his lifetime. Computers analyzed the facial features of Rembrandt's portraits and put together a typically Rembrandt human face (male, Caucasian, wearing hat, with facial hair, looking to the right). They then analyzed the typical textures of his oil paintings and a 3-D printer reproduced them, simulating brushstrokes and even craquelure. The result is convincing: It looks and feels like a Rembrandt.
This set of algorithms could be called an artificial intelligence whose mandate is to recreate Rembrandt's own intelligence. It creates a Rembrandt imagination inside a machine, a virtual personality that could be conceivably called on to keep producing Rembrandts long after the artist's death.
In this it resembles the imaginings of speculative fiction writers of the 1980s. William Gibson, in his early cyberpunk novels, described a future world in which brilliant people's thoughts and personalities were analyzed, digitized and animated by artificial intelligence so as to constitute convincingly lifelike avatars, complete with video representations, so that you could in effect talk to dead people. They are always waiting there in the computer waiting to be consulted like deathless oracles.
One could do the same these days with dead musicians. There are enough outtakes of Michael Jackson's voice in storage in the recording studios of California to cobble together an album's worth of entirely new songs; a computer could easily analyze his composing patterns and create Jackson-like melodies. It's funny, though, nobody is as excited about these computer simulations as they are by living artists.
The Rembrandt experiment is not actually an exercise in cyberpunk virtual reality or even in conceptual art. It was sponsored by a Dutch bank. It is an exercise in publicity meant to show how the bank is both technologically sophisticated and interested in supporting the arts. You can try to watch the video about the making of the painting, but it's insufferable because it's also an ad for a bank, complete with smug waffling nonsense about how banks improve the lives of people through collecting data.
This connection to marketing comes up in a lot of recent daring experiments in artificial intelligence.
Compare this super creepy one: A Portuguese company called Eter9 is working on a social network that gathers your tastes and choices into a convincing replica of your mind and can go on posting for you when you are not actually at your computer, or even after your death. You, or your "Counterpart" as they chillingly call it, will create original content, just as if you had written it. Apparently, it is one of many networks attempting to compile immortal avatars in this way.
Yes, any reasonable person would look on this idea with horror – not just because they cannot see any reason to go on posting cheery updates from the afterlife, but also because of how it perpetuates the contemporary insanity of believing that we are nothing but our social media personnas. Not to mention the fact that I might not want to have an algorithm speaking for me at any time.
People in advertising, however, being fundamentally sociopathic, see instead a great marketing opportunity. Your Counterpart is your brand: Think of how much time promoting your brand you could save if a program were to do it for you. Because that's what social media are, right, brand promotion?
What would be really great, think the publicity gurus, would be a world in which everybody is represented by an online avatar that speaks nothing but algorithmically generated, tech-marketing nonsense: Everybody would post gibberish about strategic innovation and people-driven methodologies and catalyzing next-generation solutions. And vision and empathy. Lots of empathy. We will be a world of visionary human-centred avatars, generating meaningful brand differentiation in a global marketplace. Then we really will live in the world of the dead.
Not that I have anything against talking with the actually dead. I do it all the time. That's what we do when we read their writing. I wonder what Albert Camus would think about a particular situation, and I just go and ask him – by opening up a book at the relevant chapter. He is always there to talk to me. All an AI simulation would have to work with is what he has already written.