Will AI Help You Reclaim Your Time?
People who know I’m also a writer like to ask me about AI, and I occasionally mention it when I’m speaking on productivity for one reason or another. I have some thoughts about AI, and figured this week might be a good time to drop them. Strap in!
Do we need to worry about sentient AI or AI taking over?
Ha ha, no.
If you need a longer answer than that, here goes. For the most part, tech people have absolutely no idea how the human brain works. Tech people and neuroscientists are in different silos and don’t talk to each other that much, though there are a few partnerships going on.
But your average AI team at a tech company doesn’t have any neuroscientists, just coders. And if you don’t know how a human brain thinks, how can you make a computer think?
I’m not denigrating coders, by the way. There are a lot of things neuroscience still doesn’t know about the brain and how exactly it works. Mapping the complex human brain is an ongoing project. Yet mapping a simple worm’s brain has already been done … and the researchers couldn’t figure out from the map how the worm behaves.
We can’t even define what consciousness is, much less how humans became conscious. There’s a theory that everything made of organic matter, including rocks, has consciousness. If that’s the case, then the definition is worthless, in my opinion. Humans and other animals are conscious in some way that’s different from rocks.
You could make an argument that plants and fungi are conscious to some degree. After all, trees communicate with each other through root systems, and slime mold can solve problems.
Even then there’s still some way in which humans experience consciousness differently, and science doesn’t have any good ideas (yet) on how that happens. Again, if we don’t know how consciousness arises, then how do we program computers to make them conscious?
I think where a lot of people get hung up are the terms “artificial intelligence” and “machine learning”, because most of us have certain ideas about what intelligence and learning are. However, these terms mean completely different things when you’re talking about computers.
We think of people as having some kind of native intelligence and ability to learn, which can be affected by a bunch of things. You can sit down and study something to expand your knowledge about it. You’ll learn more when you try to teach someone else what you’ve just learned, and that also makes it stick better in your memory.
A lot of learning actually happens when you’re unconscious (sleeping) because your brain is encoding the information from the day. That’s why they always tell you to get enough sleep before an exam, and also to learn over a period of time so it’s encoded in your brain.
Computers are not “intelligent” in the same way. They truly are a blank slate until some programmer writes code for them to tell them what to do. If no one writes the code they do nothing, unlike a human brain that starts learning automatically just from being alive.
And computers don’t really “learn” anything, at least not in the way that people learn. For machines, what you do is scrape a whole bunch of data and feed it into pattern recognition software, classifying things so the computer can classify them also.
How old was your kid when they could figure out a cat from a dog? Maybe toddler age, somewhere around there? How many cats and dogs do you think they saw in their lifetime? Maybe hundreds? Maybe fewer? And yet somehow without millions of images, at some point, they learn to tell the difference. It’s not entirely clear how that works.
How does the machine learn the difference between cats and dogs? Cram millions of images labeled “dog” and “cat” into its maw and run the program.
Your predictive text (on your phone or document) is a primitive form of AI. It’s the same principle: it’s been fed lots of data and it adds yours in too as you go, which makes it feel more customized. But it’s still just pattern recognition.
Significant problems with AI
You might have noticed that I stopped using AI-generated images for my posts. I did that deliberately, for several reasons.
One is that artists are not being recognized for their work when it’s AI. That’s theft. I firmly believe that artists should be recognized and paid for their work. I use photos from a free site (usually Unsplash, sometimes Pixabay), where the artists have uploaded their work and agreed to license it under a Creative Commons framework. I credit the artists on my LinkedIn posts and wherever else I can.
Most of the companies with AI capabilities scrape everything off the web, but the artists aren’t recognized or compensated. Even when the images or text are NOT on a Creative Commons license, meaning they didn’t agree that their work could be distributed for free. And yet there’s nothing they can do about it.
Secondly, why the FUCK are we asking AI to do the things that nourish and inspire real people? Why are computers generating art, when it’s a sublime experience for humans? Why does AI get to write things, when there are so many humans who write for the joy of it? (OK, writer’s strike, but that doesn’t affect business owners who are creating their own content. And for the record, I stand with the writers.)
Why does AI get to do joyful things? Because we have TOO MUCH JOY in our lives? Who thinks they have enough joy and they’re willing to give up the rest? Nobody. AI should be doing the soul-grinding, wearying BS that we humans hate doing.
In the past week I had to take my hybrid car for a smog inspection (?!) so I can renew my registration; call the brokerage company who apparently had to confirm my business address (?!) and is STILL blocking trades; get three referrals for the podiatrist because the first two are no longer practicing in their field (!!) and a bunch of other stupid, annoying, and necessary things.
THAT’S what AI should be doing: making calls and arrangements, doing errands to the extent they can, interfacing with other business systems, and so on. They’re computers and THEY should be doing the soul-sucking stuff since they don’t have any souls. (I don’t actually believe in souls, but you know what I mean by that.)
Another problem is that we don’t know what datasets the AI tools are trained on. This is how when you ask for a doctor you get a man, because so many sites on the web show male doctors and names with DR. are still predominantly male. Not to mention that the tools are used for things like policing, but the algorithms are a black box and can exacerbate racial or other bias.
Consider the Minneapolis police department, which was recently in the news for how racially biased its cops are across the board. More Black people are pulled over for the same offenses compared to white people, and they’re even worse towards Native Americans. (See article here if you want details.)
Now imagine AI using the Minneapolis police database for training. What comes out will be biased against Black and Native American people and probably anyone else who isn’t white.
These are real issues that AI companies have not solved, and it’s not clear that they’re interested in solving them.
Can AI drive a more productive day?
The answer here is “maybe”. If you’re using AI for writing, you have to fact-check it because it gets things wrong. People have found that it will cite research journals that don’t exist for references if asked. AI said my friend, a business owner, had an article in Business Week – she was very excited, but of course it wasn’t true.
Personally, I write so fast that it would take me far more time to correctly prompt the thing and then fact-check it than it does to write. I don’t recommend that people use it for generating their informative content for that reason.
I also am not particularly bothered by a blank page – I guarantee that if you just start typing something, like how frustrated you are that you don’t have any ideas, your mind will eventually latch onto something you do want to write about. I’ve done it myself a couple of times.
I also have lots and lots of ideas. Coming up with something to write about is not a problem, plus I have a content calendar for the rest of the year. If you have trouble with ideas, maybe sit down once a quarter or so and develop the ideas you want to discuss over the next quarter and schedule them in. Then when it comes to writing, you already know what your topic is.
But where AI could be helpful is getting started or generating some content and/or copywriting ideas. It’s using pattern recognition so you’re probably not getting anything new and interesting, but sometimes you just need something to get going yourself.
If you’re not sure what you should be writing about, the bot can probably give you some ideas so you’re not starting from scratch. Maybe it could generate some headlines for you, or bullet points for your article. Let it get you going and then take over from there.
Recap (tl;dr):
We don’t have to worry about the rise of the machines, but we do have to worry about which tasks AI is being programmed for and what datasets it’s using. It may be helpful to you if you’re a business owner who needs some help getting started with content, but it’s not reliable enough to use as a content generator on its own.
If content isn’t holding you back from making more money and working fewer hours, then what is? If you’re a business owner with a small team and you don’t have enough hours in the day, give us a call to see if Productivity Injection is the right solution.