What AI Cannot Do

Image © stock.adobe.com

It’s becoming tiresome, isn’t it? Every week AI is seemingly conquering new territory, doing more things, becoming more capable and more useful – or threatening, depending on point of view. High noise factor, low value because it’s mostly speculation. What about taking the opposite angle: What AI cannot do – would that be more useful?

Immediately it may seem like the same thing – speculation with a different perspective. Then, if we turn practical, we can easily find jobs, tasks, challenges etc. that AI, bots, robots, whatever the designation is, cannot do. Not now, probably never. In particular personal things, things that require YOUR eye, YOUR taste, YOUR preference, things you wouldn’t even ask your housekeeper (if you have one) to do because you’d know it wouldn’t get right. And physical things that require bodily functions, from breast feeding via a warm hand to hold, to attending to a bleeding cut or swollen eye. Or tending to the barbecue to get that T-bone or entrecôte just right.

But these are elements of life more than they are jobs. Tasks we wouldn’t want bots to do even if they could, right? What about jobs? A few years ago, renowned speaker and author Michio Kaku pointed out that ‘The jobs of the future will be what robots can’t do‘. But what is that and does it apply to your profession? Why wouldn’t robots take over and do whatever you do better than you? 

Three years ago this was an easy question. Now it’s different. Evidently, AI CAN compose music, write novels and reports and just about anything else – of surprising if not necessarily good quality. They can teach children, evaluate resumes, conduct interviews, create movies, imitate and ‘fake’ voices and pictures, drive cars, buses, trains, fly planes, analyze financial data, legal data, medical data and just about any other category of data faster and with more precision – etc. Admittedly, some of these uses have actually been around for a long time, such as autopilots and weather forecasts (as discussed in AI is predicting your future recently), but even those are getting impressive enhancements.

If you’re like me, you would argue that yes, they can do all this but is the quality good enough? Important question: What is ‘good enough’? Think about it: Is the gauge human capability or actual delivery? The latter is easy to beat, the former is not. Possibly hard to swallow, but the distance between our (human) capabilities and what we deliver has increased dramatically in recent years. No AI stands a chance against Shakespeare, Ibsen, Clancy, Cameron, Newton, Einstein, Abel etc., but do well against lousy novels, boring music, shallow (quick and dirty) analyses of all kinds, poor (or absent minded) drivers, clueless HR consultants, meaningless research etc. etc. 

Which leads to an important acknowledgement: AI is improving rapidly partly because we – you and I and mankind – are in decline. It’s hurts to admit, but we’re so busy doing ‘useless’ stuff that there is no time left to do the useful stuff right. Instead it becomes quick’n’dirty, fast, shallow, whatever is the appropriate metric for the actual discipline. Sometimes with dire consequences, such as a building collapsing, a bridge falling down, a dam breaking up, trains colliding, 10,000 misdiagnosed patients, 100,000 recalled cars etc. They are in the news every week, many – maybe most – caused by sloppiness. ‘Didn’t have time to do it right.’ Even wars have been started for the wrong reasons and based on wrong or fake data and sloppy fact checking. 

IOW, if this is our metric, it doesn’t matter that AI systems to this day have no intelligence, their (rapidly increasing) smartness is more than enough. Plus the fact that they have time to be thorough – and use it. Because they were programmed that way.

Which is my point. We’re always short on time. For generations we’ve invented and perfected all these creative technologies to assist us in order to free up time, be more productive, spend time on ‘useful’ things rather than routine work. An incredible (and accelerating) success – except we now seem to have less time than ever. Students have no time for homework, teachers have less time for each student and less time for term papers. Physicians have less time per patient, miss symptoms and botch prescriptions. Coders in a permanent rush produce bad code and more bugs, factory workers, engineers,  carpenters rush through the day and rarely take time to doublecheck and triplecheck critical points, missing errors, miscalculations, even obvious wrongs. We eat on the go, substituting real food for junk and rush. We force feed our ears with podcasts and muzak and call it ‘catching up’ – instead of letting the brain breathe, catch its breath so to speak. 

And when the evening comes, we put in a freshly charged set of airpods and listen to ocean waves, birds chirping, winds etc. to calm down and sleep — sounds more often than not created by some AI machine. BTW, I just read that the sleeping industry in the US, products and services to help people sleep better (or sleep at all), runs at around USD 400 billions a year. That’s sleep, the most natural thing in the world. The list goes on, feel free to add your own favorites…

This sounds like slow, self-administered suicide because it is. The new generation of (smart) tools and helpers can either accelerate the process or help us break the downward spiral. It’s our choice, and the big surprise shouldn’t be the capabilities of the new technologies but why we keep asking the wrong questions. 

We should know by now that more tools will not give us more time. It has never happened before, it will not happen this time. More ‘slow time’ – which is what we’re talking about and what we need – isn’t something we can buy or invent or put in a battery or a bank account. More slow time is a decision. More slow time (or just ‘more time’ if you like) provide the opportunity to think, to reason, to breathe, to rediscover sanity and common sense (which was once a given…), to understand, make right choices, to live.

AI cannot do that. And AI cannot help us do that. On the other hand, if we – you and I – decide to make available more slow time, time to be and time to think, we can make use of AI in ways you’d never otherwise think of. Why? Because we open the door to our own capabilities. How’s that for a next step? Think about it: How would YOU use the incredible smartness, the amazing opportunity AI may represent? Save time or create a future?

Related posts:

1 Trackbacks & Pingbacks

  1. What AI Cannot Do (Part II) – mindset3.org

Leave a Reply