As artificial intelligence becomes ever more sophisticated, should we be scared of the inevitable changes… or optimistic?
Being victimised is a popular pastime these days. From our politics to the content of our Netflix series to imminent environmental collapse, there’s no end of ways to make you think you’re on the cusp of being wiped out by some threat. A particularly popular front runner in the contest for ‘most awful thing that’s going to happen to us’ is artificial intelligence – particularly when married in the imagination to sophisticated robotics.
I’m proud to say I can gloom it up with the best of them. Ask me where we’ll be in ten years and I’ll spin you a tale that will make Cormac McCarthy’s The Road look like The Muppet Movie. Fantasies of out-of-control artificial intelligence are jet fuel for dystopian daydreamers like me. But why is AI so scary? What are we worried about?
Well, first of all, humans have a lousy record of using powerful technologies without inflicting harm. Whether it’s blowback from fossil fuels, or antibiotics training bacteria to become deadly super-blobs, our big tech advances often have negative sides we are loathe to confront until they threaten to wipe us out. For every great tool we discover, a shadow also appears. (In the US, the leading cause of death among children is car accidents, with firearms the second.) Artificial intelligence seems to be a perfect storm for these anxieties, especially given that it’s designed to magnify the human will. Which is a great thing for me, but in your hands? No way.
direct threats can always be neutralised by taser-spitting robotic quadrupeds
There is a deep-seated suspicion that the goal of AI research is to supply us with synthetic slaves that will do what we want, when we want, better than we could ever hope to do. We are also aware AI and machine learning are technologies that ought to get exponentially more sophisticated over time, with AIs training AIs. The ability for AI to alter how we live could accelerate very rapidly, far more rapidly than we can adapt. With AIs already able to outwit humans in many strategic settings, there is a growing dread that AI will not be responding to us anymore, but that we will be responding to it. In fact, it’s possible this is the case already.
Aren’t we are also afraid of AIs that are programmed for self-preservation? (Robot marketers will call them ‘robust’ and ‘self healing’.) Afraid they will (rightly) conclude humans have the power to literally pull the plug on them and so need to be neutralised. This might not involve extermination – just the production of some especially addictive entertainment that keeps us immobile. And direct threats can always be neutralised by taser-spitting robotic quadrupeds that can run as fast as a horse.
As if robot overlords weren’t enough to worry about, there’s also a sustainability problem. (There’s always a sustainability problem.) AIs are going to require a huge amount of energy. In the last issue of FEED, we looked at the surprising amount of energy that video consumes worldwide. It’s thought the computing power required by AIs could be substantially greater.
At a recent conference, Gary Dickerson, CEO and president of semiconductor manufacturer of Applied Materials, predicted a worst-case scenario at a recent conference in which AI computing could make up a tenth of the world’s electricity use by 2025. That is, unless new, more sophisticated chips are made available – chips which (one supposes) Applied Materials could supply. 10% could be an overestimation – or not – but there’s no doubt that AI is going to demand a lot of electricity. Just think of the power required to coordinate cities full of autonomous vehicles and fleets of delivery drones 24/7.
A natural part of AI in the future should be the monitoring of its own resources, optimising for efficiency in ways that humans could never calculate by themselves. But given the truism that demand always exceeds capacity – we tend to always spend everything we have – it’s a near certainty that energy savings created by AIs will be used up by even more energy-saving AIs.
Then there is the assault by AIs on the human mind itself – on human autonomy. We can barely outwit the simplistic algorithms that keep us glued to social media platforms. It won’t take too many years for AIs to bend us to their will – or their owners’ will – in ways we can’t yet imagine. Two years ago, Google’s AlphaGo computer beat the world’s top Go player; a dark day for AI-phobes. An AI of the future, whether it works for a security company or an ad agency, is always going to be 50 moves ahead. Sure, it looks like a simple ad for a concert, but is it really just the AI setting me up to get turned onto a band whose music it will try to sell me on a new social platform, which it will use to absorb some new data about me? Is Tinder really just an AI-driven eugenics program? The possibilities for paranoia are endless.
So yes, if you want to worry about artificial intelligence, I won’t stop you. There’s a lot to worry about. But…
What if (and I know this is crazy) we had some faith in ourselves to intelligently answer the many questions AI poses? What if we assumed the very best for the world of AI? The idea that more technology leads to an improved quality of life has long been disproved, but maybe the AI future has much more positive potential than we imagine? Maybe we are entering a golden age of problem-solving? Maybe your AI future isn’t going to be a dystopia. Maybe it’s going to be a utopia.
So, let’s imagine it. Your AI future. In the AI utopia…
The playing field will be levelled
No one is going to stop you from going head to head with someone at your local squash court, but in the future AI will open up strategic options both in business and in personal life that were previously only available to the naturally clever or those with specialised degrees. We will all become fantastically self-sufficient at using resources. Rather than flipping through books, YouTube videos or calling contractors to try to fix a home a plumbing problem, AI will collate and organise specific solutions for your specific situation and, depending on how much information access you’ve given it, will probably have anticipated the issue long before it becomes a problem.
Maybe your AI future isn’t going to be a dystopia. Maybe it’s going to be a utopia
Education will take on a whole new framework. Facts and information will be available even more readily than they are now, refined and tailored for you needs. But, more importantly, AIs will work with you like coaches to improve your use of that information, readily integrate new data and inputs into your life, and give you feedback about your progression to your goals.
Businesses will have access to financial strategy AIs that can forecast many months or years ahead – utilising mass volumes of local and global economic data. Business finances will not be focused on maximising profit, but on maximising value. AIs will be able to suggest ways to leverage even small amounts of capital for maximum results. Bigger will not always be better. A small company with a smart AI will be able to accomplish far more than a big, well-funded enterprise working without AI.
Products and services will be directed to where they’re wanted and needed most
Rather than creating a product or developing a service and then hoping you can find an audience for it, AI will be able to crunch huge amounts of consumer data to get those offerings directly to the people who want and need them most. This will aid both the producer and consumer. Consumers will be able to fine-tune what they want at any given moment – healthy vegan food or something decadent? Content to stretch my brain or just something for numbing out? And producers will already have a clear idea of who exactly wants their product and what the potential demand is for other similar products.
With such transparency in the market, advertising will be transformed. And may become irrelevant altogether.
Content will not just be built by AI, content will be AI
We are entering a phase of content production where AIs will increasingly contribute to the production of content. AI-assisted editing and capture are already popping up here and there, but in a generation AIs won’t stay in below-the-line jobs. In the future, content will be indistinguishable from the AI. Creatives will be people who hone, coach and prune the AI to produce a whole array of engagement experiences and community and monetisation options. Consumers will interact with the AI through various means – video, audio, games, discussion, experiences. The AI will be a multi-faceted entertainment, teacher or experience portal, which can interact with and respond to its consumers in real time, based on the parameters cultivated by its coaches. The entertainment AI of the future will be a digital garden, always changing – alive and not entirely in the control of its caretakers. When you buy content in the future, you won’t be stuck with buying an individualised story or work, you will be buying access to an experience that is produced by and takes place entirely within an AI.
what’s clear is that artifical intelligence itself is never going to be the problem
It’s possible that businesses of all kinds will operate this way, as well. Inc, LLC, LLP or PLC will be replaced at the end of a company name with the types of AI on which the company is based. All businesses will be AI-enabled businesses, and most will be AI-centric businesses. In many cases, the AI and the business will be the same thing. Directors, employees and specialists will simply be attendants and benefactors of that specialised, cultivated and trained AI.
AI will be your matchmaker
Rather than crudely sticking together people based on simplistic data points about personality, likes and dislikes, AIs of the future will be able to find connections between people not based on who they are (or have been), but on what the potential is in their interactions. A sophisticated AI, through accessing large amounts of past data and running probability simulations, will be able to see the possibility in the spaces between people. Rather than Person A + Person B = A nice date and possible smooching, the AI will determine Person A + Person B amplifies the possibility of an interaction with person C, which causes Person D to bring all four of them into a cooperative endeavour to create something new and marvellous.
This needs a certain amount of faith in the AIs that put together pieces in a puzzle that seem nonsensical to us, but which, over time, will yield results we could not have predicted with linear thinking. These new AI-assisted social groups will not be the self-reinforcing rabbit holes social media has forced us into today. Instead, they will be open-ended systems that encourage discovery, possibility and redraw road maps in real time continuously. Whereas today’s AI might read something outside the target as a problem, a new Exploration and Discovery AI will encourage the import of new experience and data into the larger system. Relationships and groups will be environments for discovering new things, not to keep getting what you’ve always got.
AI will enhance individual freedom and autonomy rather than restricting it
We fear Rise of the Machines, in which we are dominated by AI and our freedom and power of decision taken away from us. But what we are seeing in this negative vision is only a half-baked idea of what AIs are capable of. The principle purpose of AI is problem-solving. What we might discover, when problem-solving becomes ubiquitous, when AIs are able to outthink even our most problems, is that… oops… we still have problems. The human mind is an endless problem generator. That’s kind of its job. We love solutions, we humans, and you can’t find a solution unless you have a problem. It’s conceivable that an endless array of problem-solving AIs may turn us back on ourselves. With great power comes great responsibility, says Spidey’s Uncle Ben. Is it possible that the great personal power future AIs bring us will ultimately encourage us to think about how to use it best? Will AI paradoxically encourage us to be more responsible? More reflective?
Electricity is something we have to work to generate 24/7. Our lifestyles – our lives – depend on its continuous flow. But few people could seriously make the claim that electricity has reduced our autonomy rather than enhancing it. AI is going to make you more independent – if you want to be – and give you more choice – if you want it.
Maybe we can put aside our fears of the AI apocalypse just for today, and have hope we’re not entirely hopeless 2001-style apes. Maybe we can achieve some good with our tech. It’s a shift that can be very helpful, as a thought experiment if nothing else.
What’s clear is that artificial intelligence itself is never going to be the problem. AI tools really do have the potential to move mountains that previously wouldn’t budge. The area of concern is that AI is going to be aimed at the things that humans think are problems. As we’ve discussed, a human’s ability to distinguish a real threat from a phantom, or to distinguish a craving from a legitimate need, can be very flawed indeed.
To get the most out of AI, we won’t need smarter machines, we’ll need smarter people. STEM training is great, but what is going to enable us to use technologies for the benefit of all is not better technicians, it’s better philosophers, artists and thinkers. To get the most out of AI, we’ll need a whole generation of people who know not what AIs need, but what people need
This article originally appeared in the September 2019 issue of FEED magazine