As with all things, AI is not black and white

Three years ago, AI was still barely on the radar of the general public.

We were all recovering from the onslaught of Covid, and the vast majority of the world was just getting back to normal operations when ChatGPT entered the scene with a big splash. I remember all the buzz around it on Twitter.

It was presented as the best thing to hit the internet since the cryptocraze and meme stocks.

On social media, you fell into three camps:

• the skeptics

• the early haters

• the early adopters

I was mostly the first two, and with good reason.

As an engineer with an extensive background in electrical and computer engineering, it had crossed my consciousness a few times throughout my college career. I wasn’t all that interested in jumping onto the bandwagon as I had seen every other technology fad come and go as well. I took my time to poke at it in image generation as well as feedback, idea generation, and prompt engineering.

Now that it has had a few years to simmer, I have a few things to say about it…

It isn’t a silver bullet

Remember when the TI-89?

That was the smartphone of the mid 90s. Every kid had to have this graphing calculator, and it actual had enough memory to load games on it as well. It was the closest thing to a Gameboy that you could actually take to school with your parents being none the wiser (so long as you didn’t show them all its functions).

AI is now there, but just because you can use it for everything absolutely does not mean you should.

Long-term usage of it is the same as what happened with people and calculators. In the most extreme cases, most people can’t even do simple addition without whipping out their phone app to check. Some kids can’t even count change at the register…

It’s easy to abuse technology and use it as a crutch for everything.

There’s nothing wrong with leveraging its vast abilities for productivity purposes, but when you resign yourself to only handling special cases of mundane thinking, you risk accelerating the Ebbinghaus curve on a few levels. It’s the rate at which you forget things that you don’t use or haven’t touched on a mental level, and it was something I came across with my interest in learning multiple languages at one point.

We’re not computers.

AI will never need to maintain its memory in the way that we do. You run the risk of becoming more reliant on it over time if you aren’t careful. As AI agents become more popular, the app store is becoming inundated with many different types of them.

If you weren’t hooked on your phone before, needing AI for everything will make it impossible to function without it.

It can’t replace you. Ever.

You’ve seen the hype.

People are using it to create virtual influencers complete with AI-generated pictures, AI-generated social media posts, and even AI-generated chatbots. You can resurrect famous individuals from the past and interact with their avatars to ask questions and interview them based on the content they left behind.

These are echoes at best.

The thing about AI is that it relies on what it can access and train on across the internet. Legal issues aside, it is limited by extent of human experience and knowledge that is present. It can’t take two things and create something entirely new. It can only place them together in a strange collage-like illusion. This will change as it improves, of course, and video generation is already significantly better than the days of watching an AI-generated Will Smith eat spaghetti.

Image generation is already rampant.

I’ve used it for my own writing in the past, but it occurred to me today that it could become a detriment to my readers. As AI is getting better all the time, it’s still fairly easy to see when an image is generated or authentic. Here’s the question, though:

If readers see AI-generated images complementing my writing, what are the chances that they might think my writing is the same?

It’s a slippery slope like when you are caught in a lie or having an affair. Once the trust is breached, there is no going back. The teacup has shattered, and the cracks will always be there no matter how good the reconstruction was. Building trust with your audience is one of the most important things that a creator has to do regardless of the medium they use.

The people that I love to read established their credibility before they started using AI for their images. Perhaps I’m making something out of nothing, but it is something that I wanted to put forth for consideration.

Trust is in very short supply these days.

You can’t believe everything you see on the internet, now more than ever. Don’t fall into the trap of thinking that you can use it to capture your style and tone and voice and likeness to generate content for you. It is absolutely possible to do now, but this goes back to my previous point.

It can only represent who you were.

It won’t represent who you become.

Use it for these things

I now enjoy using AI for virtual feedback and for idea generation based on things that I come up with. The weekly topics that I write about now are based on these results.

As an engineer, I can see where AI shines, and it already is making a huge impact in science and technology. Protein folding algorithms have vastly benefitted from it. Virtual design optimizations are now at a fraction of the time it used to take. Computer circuit layout and iterative upgrades are faster as well.

AI should be used for productivity, and knowing when to apply it requires critical thinking skills.

Tread carefully when using it for learning and self-expression. In academia, plagiarism is now much less of an issue than AI-generated responses in essays and term papers. Computer science is one of the worst places to use it because programming languages require repetition and iteration.

The most powerful way to demonstrate proficiency and knowledge is with an interview or an oral exam.

This is where the most challenging aspect will be for future generations. Someday, our descendants may all end up in rooms like the one in 2009’s “Star Trek” reboot. I actually played around with this feature on ChatGPT by asking it to challenge my knowledge of US history and quiz me for a bit. It was fun, but I made sure to double check my responses with a textbook as well as searching for it in the traditional way.

What about you?

How are you enjoying your AI experience?

What are your thoughts on AI and its potential?