• by Gord Hotchkiss , Featured Contributor, Yesterday

I have dipped my toes a little further into the deep, deep waters of AI (which this week promises to get deeper with the recently announced OpenAI IPO). Here, then, are my updated thoughts on artificial intelligence.

For the past few months, I have been doing post-production on a documentary. The footage has been shot over the past six or seven years, and I have been wading through hours of interviews and other footage, writing the script, building edit sheets, and -- for the past month -- actually doing the edits. In the process, I have tried to use AI judiciously. My rule of thumb has been this:  If I’m going to use AI, don’t be obvious about it.

Because I use Adobe tools for editing and some special effects, AI assistance is now built into almost everything I do. It lurks beneath the hood on pretty much every menu item and tool selection Adobe offers. This is not the type of AI I am referring too. This built-in AI is what I call covert artificial intelligence.

I am referring to overt AI, where I have made a deliberate decision not to use the limited processing power of my own brain and have unleashed the AI Kraken to do my bidding. In my case, that has involved gathering information for the script, creating soundtracks, some video special effects and some other technical aspects.

I did stop short of using AI for some special video effects that created an uneasy feeling in my gut when I saw them. Something told me I had stepped over the threshold into the realm of excessive creepiness.  When you’re dealing with people’s memories, especially of loved ones since departed, you have to tread carefully. Just because AI lets you do something, doesn’t mean you should.

But here’s my point. AI couldn’t have made that call. It couldn’t replicate my intuitive understanding of my audience, even to the point of anticipating reactions from specific individuals whom I knew would be watching. AI can’t make gut calls about human consequences.

In that, it bears a resemblance to Phineas Gage.

Unless you’re a neuroscience geek, you may never have heard of poor Phineas, but it’s one of the most famous case studies in science. On Sept. 13 1848, Gage was working with a railway blasting crew in Vermont. Due to an unexpected explosion of blasting powder, an iron crowbar was driven through the left side of his frontal brain. Miraculously, Gage retained much of his memory, language and reasoning ability. What was impaired was the ability to make gut decisions, using something neuroscientist Antonio Damasio called “somatic markers.”  

It’s exactly these instinctual abilities to predict the responses of others that AI lacks. While it may be “intelligent,” it is not necessarily “wise,” as wisdom requires judgement. It requires the ability to foresee reactions and to adjust your decisions accordingly.

When we look at the intersection of creativity and AI, it’s this human aspect that we shouldn't be too quick to surrender to technology.

Take the writing of this post, for example. I did use AI to help me gather my thoughts. But the writing itself is -- for better or worse -- all me.  

I have been writing for pretty much all of my life. I have my own distinctive human voice. My grammar isn’t perfect. I am Canadian, so my editor (the incredibly patient Phyllis Fine at MediaPost) has to “Americanize” my spellings.

I started my career writing for radio, so my sentence structure can sometimes be described as rambling. But all this accurately reflects who I am. If I used AI to actually craft my words, I would have removed myself from the loop. Even as I write this text, my computer screen is littered with blue dotted underlines suggesting there may be better ways to word things.  I’m ignoring almost all of them.

I don’t know where the right AI-human balance is. Even if I did know, it would be a moving target. The things AI can do today are a quantum leap ahead of what it could do even a year or two ago. But I don’t see AI gaining the ability to actually feel anything anytime soon.