The real risk is an AI that wants us to be happy

Among all the prophecies of nuclear wastelands littered with human skulls crunching under the feet of murderous robots, is there room for hope in the AI-dominated future?  

Ask anyone who the ultimate male Hollywood screen legend is and the debate will rage forever. Humphrey Bogart or Clark Gable? Jimmy Stewart or Marlon Brando? Robert De Niro or Al Pacino?

All of the above would probably get a worthy mention on anyone’s list. Someone who I’m pretty sure wouldn’t, though, is the man who turned insipid, Austrian monotone into one of the most valuable (and parodied) commodities in Hollywood history.  

One of my absolute favourite movies of all time is the seminal—yes, seminal—Terminator II, starring Arnold Schwarzenegger. Its final scene stands as one of the most tear-jerking moments I’ve ever witnessed on the silver screen.

As our hero—or Cyberdyne Systems model T-800 to give him his full title—lowers himself into a vat of molten steel, sacrificing himself to protect the humans he has learned to love, my eyes never cease to moisten.   

For my money, T-800’s line to John Connor, “I know now why you cry”, is up there with “Frankly, my dear, I don’t give a damn” and “I’m gonna make him an offer he can’t refuse.”

A psycho-robot-killing-machine that discovers empathy! What could be more heart-warming in this age of existential AI-hysteria than that?

For all this talk of AI taking over the world and establishing a dictatorship of—what shall we call it?—“the technokariat”, I thought it might be wise to consider what this [delete accordingly: utopian paradise/dystopian hell] might actually look like. 

AI learns from us, so it makes sense that our putative Overlord Supreme looks at its human creators for the perfect form of absolute rule. There are plenty of historical models for it to ape. 

The blueprint for establishing a dictatorship is pretty clear. First, exploit a crisis, then seize control of the media and armed forces, follow that with a quick purge of your enemies, establish a personality cult, and—bob’s your uncle—you’ve got your very own Death Star. 

An idiot’s guide, basically. Simple enough to condense into a ten-second Instagram reel. And I’m pretty sure with its superintelligence and access to the world’s entire technology grid, AI would make a better fist of tyrannical rule than any of its human predecessors.  

But, if AI truly is destined to be our Imperial Majesty, will it really lean towards the Emperor Palpatine end of the spectrum? Or might it edge closer to, say, Gandalf? 

Among all the prophecies of nuclear wastelands littered with human skulls crunching under the feet of murderous robots, is there room for hope in the AI-dominated future? Will our new rulers really pursue this scorched-earth agenda or might they have something else in mind for us? 

“Power tends to corrupt and absolute power corrupts absolutely,” so said Lord Acton. But would an empathetic AI be as susceptible to corruption as you or me? Would it possess the same human vices—greed, jealousy, paranoia, cruelty to name but a few—that have dragged our species into the gutter for millennia? 

Let’s give it the benefit of the doubt. Let’s imagine a future with AI as benevolent dictator of the world. AI could distinguish the brightest and best more effectively than any exam or education programme. Imagine professions filled with the best people, singled out at an early age and nurtured to their fullest potential to achieve the greatest good. 

Think of the kinds of pioneers and innovators AI could unearth. A true meritocracy in other words. Sure, we wouldn’t need mathematicians or theoretical physicists anymore—not in the sense we have them now anyway. But imagine the work humans could do alongside AI as it powers us to ever greater heights. 

Already AI has helped solve multiple long-standing mathematical problems. But that makes humans redundant, right? In one sense, yes. Yet perhaps the human role shifts towards a more advisory one—determining which problems need to be solved.

And how about security and corruption?  Our AI protector could root out those unsuited to wielding power and influence—the psychopaths, the narcissists, the crooks, the liars—and bar them from ever holding public office, practising medicine, leading armies or any other profession where human lives are at stake. 

Corruption, incompetence, negligence—erased like a wiped internet cache. The closest the criminally negligent and incompetent would get to office would be gazing at it from the depths of social obscurity—if not through the bars of their prison cells. 

AI is already used in criminal profiling, predictive policing, criminal risk assessment, evidence collation and more. Although such systems are imperfect and can produce false positives, once they are refined, what might they achieve? What if they are pushed further and asked to pre-empt criminal behaviour before it manifests? 

Interventions from teachers, social workers and family members could occur long before individuals have ventured over to the dark side. Think of the number of crimes this could prevent, the lives saved, the families spared grief. No longer would we castigate law enforcement and social welfare bodies for failing to act sooner, missing red flags, ignoring warnings.     

And what about dating and relationships? Could AI determine our perfect partners years before we’ve even met them? Our soulmates found for us? Everyone paired off with their ideal mate? No more toxic relationships? No more cheating spouses? No more loneliness? 

Perfection achieved. 

But then what does perfection look like from the bottom rung of this meritocratic utopia? What if I’m identified as a potential miscreant before I’ve so much as shoplifted a Mars Bar? What if I’m an ‘undesirable’? An ‘unmatchable’? What if I don’t fit in with the ethos of this perfect society? What if the problems I want AI to fix aren’t the problems AI thinks need fixing? 

What if the problem is me?

And then another deeper question, too: Even if I find myself at the top of the tree, paired off with the love of my life, and thriving in this brave new world, how much of this wonderful life did I actually choose? Am I happier now that everything is decided for me?  

I think it comes down to a fundamental conundrum: Would you rather have every desire and need catered for at the expense of your freedom to choose? Or would you accept poverty and uncertainty in exchange for the right to do as you please? Do we want to be happy or do we want to be free?

So, assuming AI does turn out to be a sage, omnipotent ruler, it probably knows the answer already. We humans make mistakes, but more often than not we prefer it that way. We would rather follow our passions, screw things up along the way, learn a little, make some more mistakes, and revel in the joyous mess of it all.

That sounds far better than having perfection spoon fed to us. As Paul Dano put it in Little Miss Sunshine: “You do what you love, and fuck the rest.” 

So, perhaps the disappointments, the fuck-ups and the emotional baggage are the very price we pay for freedom. And perhaps the price is worth paying.