Artificial intelligence (AI) is everywhere around us: it makes medical diagnoses, writes magazine articles, creates works of art, makes all of our objects “smart,” and responds to our most difficult questions. All of this with an undeniable efficiency. Autonomous cars will avoid 90% of accidents on the road. Algorithms detect cancer faster and more precisely than a doctor.
In theory, AI is also impartial, avoiding gender and race prejudice for judicial decisions. Leaving us only tedious and time-consuming jobs. It will predict and avoid financial bubbles and stop flawed recruitment. It will develop perfect products and make things that were once complicated easy. A dream
No more work, but what else can we do?
Instead of permanently daydreaming, we might end up wanting to wake up. Take working for example. Robotization and automation will eliminate between 14% and 47% of jobs according to different studies. Good news for those who have a boring and horrible job, but do lawyers, surgeons, artists, and sports commentators really want to be “free” from their job? And what are we going to do with our “freed up” days and all our day-to-day tasks?
Getting groceries and do the cleaning? Robots will take care of that. Plan our next vacation? An app can do that. Go shopping? Amazon will have already found us the clothes we need. Write a history book or storybook for children? Not a chance with “algorithm writers.” A day powered by AI might end up seeming very long indeed.
This film is so predictable
Entertainment also risks becoming bland. Several startups are already offering help to Hollywood producers. For example, Vault can predict a 75% success rate of a film by just reading the script and analysing hundreds of thousands of parameters. Other startups pitch to studios by demonstrating what type of superhero, plot, and the ending has the best chance of captivating viewers.
Television is also in on the act. For the series “Orange Is the New Black,” Netflix precisely analysed data to ensure future success: the themes of the prisons and crimes, the black comedy and the presence of a charismatic female lead. This helped reinforce the validity of the format before production began.
“It’s my worst nightmare,” the scriptwriter Ol Parker says, panicking already. “It's the enemy of creativity, it can only replicate what has already worked which can only lead to boring films that are all about the same thing.” A third-rate series like Shark Attack or French-Fried Vacation can be just as funny as a polished comedy produced by algorithms and statistics.
Remember when the trains used to be late?
The randomness of human error is what is missing from AI. Would we get the same pleasure of watching a football match refereed by a robot? “Rules allow it, but [...] what also interests us is the power to say that the referee is wrong, rather than that the player cheated.
A completely rational world would be totally boring,” proclaims Pierre-Henri Tavoillot, president of the College of Philosophy and professor at Sorbonne. What are we going to talk about at the water cooler, if we can’t complain about trains running late or grumble about the head of HR who messed up our payslip again this month?
In 2016, an Australian found an extraordinary prehistorical site by taking a wrong turn. Guided by GPS, he would have driven straight by. In 1928, the Scottish researcher Alexander Fleming discovered penicillin by “accident”. When he came back from his holiday, he found that his abandoned Petri dishes had been colonised by mould.
Clearly, if he did the same thing during the time of automation, an alert would have told him to clean his material before leaving to avoid this “error”. Velcro, corn flakes, tarte tatin, post-it notes... Some of the best inventions are because of clumsiness or human error. And these concepts are completely outside the parameters of artificial intelligence. As such, Innovation itself is in danger.
Would Ferrero have launched Nutella if they had kept in been analysing nutritional facts, regulation risks or the price of ingredients (the price of hazelnut fluctuates a lot)? Furthermore, with our refrigerators ordering our groceries automatically based on our previous purchases and our predetermined tastes, how hard will it become for manufacturers to launch new products?
The (too) perfect partner
On an imaginary day in 2030, the American consultant Dan Clay describes his life saturated by applications and services. No more awful dates and hours going through profiles on dating websites. Artificial intelligence immediately found him his completely compatible alter ego. “I eliminate candidates by the slightest detail that could bring about a possible incompatibility (could I really marry someone who wears orange?),” he says.
In the end, the dates that he goes on end up being terribly boring. “We have so many things in common that we don’t have anything else to say to each other.” Our alter ego seems more like a robot than an ideal partner who can get angry with us, discuss choices, ruin our football nights, or spark a fit of laughter in seeing our mismatched shoes. We will hardly feel like talking with our smart speakers.
On paper, AI will make a seamless world, without surprises or bumps in the road. But what would happiness mean without hardship? What would success mean if we don’t fail first. What can we be glad about if all our wishes came true? More so than being boring, a totally predictable future is also scary: the American startup, 23andMe, claims to estimate the risk of breast cancer, Alzheimer’s, or Parkinson’s disease according to genetic profiles.
Aspire Health has created an algorithm that predicts your probable death date, used officially to determine which patients should benefit from palliative care first. But do we really want to know how and when we are going to die?
