Can robots and AI be our friends? It looks like we won’t be having AI friends any time soon.
Japan has a problem.
Well, it has many, like most countries, but there is an acute one, and that is the ageing of its population. Currently, a quarter of Japan’s population is over 65 years old, and by 2050 that segment of the population is estimated to reach a third.
Europe and other regions worldwide are also ageing rapidly, but Japan is ahead of the rest. Add to that historically low levels of immigration, and the problem is compounded.
Who will take care of the elderly? Who will do the work to sustain the economy?
Enter the robots…
Yes, you read it well. Robots may be the solution or part of it.
We have discussed automation extensively already here, but today I would like to focus on another area of robots and AI that we haven’t discussed so much in this blog (except maybe in Love in the Age of Machines).
Can robots really take care of human beings? Can they understand our emotions and, be caring, and show empathy?
Can they be our friends?
AI friends in fiction
In the aforementioned “Love in the Age of Machines,” we mentioned the film Her, where Joachim Phoenix falls in love with an AI which understands him better than any human being possibly could.
The film is at the same time funny and sad. Phoenix’s character is naïve and melancholic, but we all kind of feel for him and understand how he falls for the AI. At the same time, it’s quietly horrifying because he is falling for a machine, and we find that naturally repulsive. It feels wrong.
Her may sound like science fiction, but it’s closer to reality than you think.
Some existing AI friends
There are already similar solutions out there, like, for example, Replika, your AI friend. The idea of Replika is that you build your own AI friend to your liking, and you talk to him or her, building a relationship with an AI that is there to support you and help you and who understands you better than anybody else.
The problem is it doesn’t seem to work. Some of the responses seem to be clunky or awkward, and the AI friend doesn’t seem to be good at listening and understanding you. That doesn’t mean that the technology won’t improve and eventually get there one day, but it isn’t ready yet.
Another AI company, Affectiva, is developing AI that understands human emotional states based on facial expressions, voice, and other physical cues. Again, this is still far from being fully developed. They cannot yet compete with human beings, who have an advantage of millions of years of evolution, in understanding human emotional states. They might get there eventually.
Yuval Harari, the author of Sapiens, believes AI will know our desires and understand our emotional states not only better than other human beings but even better than ourselves.
When that happens (if ever), what is to stop AI and robots from becoming our best friends? At the end of the day, what do you want from a friend but understanding, shared values, and forgiveness? Can’t a machine provide those?
I don’t think it can, at least not for the time being or the foreseeable future, but I’m known to be often wrong.
Caring is for humans
In AI Super-Powers: China, Silicon Valley and the New World Order, Kai-Fu Lee believes that AI will replace human beings in most jobs as we improve technology and everything gets automated. The only jobs that will escape this lurch towards “the automation of everything” will be those requiring human care and love.
As he says:
“For all of AI’s astounding capabilities, the one thing that only humans can provide turns out to be exactly what is most needed in our lives: love.”
Kai-Fu Lee believes that human beings will be taking all the caring jobs that require emotional connection because robots and AI won’t be able to replicate that connection correctly. Even if they were, we would still prefer to be cared for by a fellow human. Jobs like nursing, taking care of the elderly, the infirm or children, providing social assistance, etc., will be carried out by humans.
This makes sense, and it’s easy to agree with, but what happens when there aren’t enough humans to take care of the elderly and the sick, like in Japan? There, robots give needed company and solace to some lonely elderly, who seem to appreciate them. It seems to be working.
I agree with Kai-Fu Lee on this one.
Are friendship and emotional bonds the remit of humans only (and some animals)?
I can’t imagine myself creating a meaningful emotional bond with a machine. I think it is an absurd and crazy idea, even abhorrent. How can a piece of metal and silicon chips understand me, and how can I be emotionally attached to it? Aren’t emotions what make us human?
I am probably biased, though, and I’m seeing this from my own personal belief system. I
didn’t grow up with AI and robots next to me, so I’m not used to them, and I don’t think they are very good (yet) at simulating human emotions and striking up pleasant conversations. They are still prone to error (try to have a conversation with Siri and see what you get out of it).
However, if I were old or sick, and I had nobody to talk to, perhaps I would appreciate being listened to by a machine that answers back, gives me company, seems to understand me, and doesn’t judge me.
It is easy to judge this from my current position and think it ridiculous, but I guess I would look at it differently if I were lonely and a robot was my only companion.
Artificial Intelligence isn’t so intelligent
In her book Artificial Unintelligence: How Computers Misunderstand the World, Meredith Broussard explains through real-life examples how AI and computer programming work and how it isn’t that intelligent after all.
It is in our nature to provide some sort of agency, or even attach a personality, to some things depending on how they behave. We are starting to believe robots and computers can be intelligent and think a bit like us, but they aren’t. AI works via statistical inferences based on troves of data.
DeepBlue can beat Kasparov at chess, and AlphaGo can do the same with the Go world champion, but they don’t know that they have won. They don’t know what winning means. They don’t even know that they exist. They don’t know anything apart from what the best move out of millions is to win a game.
The same thing happens with GPT-3, the AI that can write almost like a human. It can write something that can pass as something written by a person, but it doesn’t know what is writing or what it means. It just gives an output that is likely to do its job based on the billions of data points it has analysed.
I agree with Broussard. Artificial Intelligence isn’t very intelligent today. That doesn’t mean it won’t be one day.
How far away is that day? Years, decades, centuries? Nobody knows.
If AI today isn’t sufficiently intelligent to do some things that are rather basic for us, it is even less emotionally intelligent to interact successfully with human beings.
A non-sentient AI that doesn’t even know it exists cannot understand all the complex emotional phenomena going through human beings. They can pretend they do, and they can certainly fool us, but that’s enough to build a meaningful emotional connection.
We are social animals
Human beings are social animals.
We crave interacting, connecting, and bonding with other human beings. This can often be extended to other animal species like dogs, cats, and other pets. Can it be extended to robots?
Robots may have their place responding to some basic needs, like providing companionship to lonely elderly or acting as the loving partners of even lonelier young people.
It is too early to go beyond that and offer anything resembling a human connection to people in normal circumstances. They are too clunky, cold, and mechanical, and they are still prone to many errors in their conversations. Most of us still have difficulties relating to them.
Who knows, in a few years, I might be looking forward to having a chat with my dear friend AI and sharing with it what’s bothering me at work and how I feel, but I think it will still have to pass quite a bit of time and the technology will have to improve considerably.
Weirder things have happened, and technology is advancing at an accelerating pace, so it might happen earlier than we think. Until that happens, I’ll continue enjoying the company of my dear flesh-and-bone human friends.
Friends, I love you all. I wouldn’t replace you with a machine for anything in the world, at least not yet.
In the future, who knows…