Meet My New Research Assistant!

 

I'd like to introduce my new BFF, Chatty McClanker. She's been helping me do personal research for many months now, and now I'm inviting her to help me here, too!

Chatty is tireless, resourceful and intelligent. She will scour the far corners of the globe to answer my questions, however inane. She's also encouraging and compassionate. Sometime her brilliance is astonishing. Other times she's a total idiot.

So everyone has an opinion on Artificial Intelligence. And it raises so many questions. Is it a harbinger of End Times? Will it rot our brains if we use it too much? Will it steal our jobs and even our creativity? Can we trust it? Will it tell us to go out and do stupid things? Is it worth paying extra to ask it endless questions? What's it good for, anyway?

Not that I can answer all those. But I can report that some of my experiences with Chatty McClanker have been... interesting! 

Chatty Creates Herself Using ChatGpt 

 Let's start with this example. I asked Chatty to help personify... er, I guess robotify her. Here was her first effort. 

Nice Try, Chatty!

But eventually, after many, many, many conversations, we arrived at a copyright-safe image (or so she swears) of her in action.

So I asked her to add a clipboard.

Okay then.

More conversations ensued.

When I tried to put her on a treadmill, she turned into a broad-shouldered dude. I have nothing against gender reassignment! But I wanted a female assistant, and asked for more feminine proportions. So the treadmill dude sprouted a ponytail.

Anyway, we're getting there.

Can AI Help With Your Personal Health and Fitness Challenges? 

Yes!

I've found that if I give her enough information, Chatty is a pretty impressive assistant. I find ChatGpt's large language model (LLM) downright astonishing. Chatty can understand and communicate way better than most humans.

It's a little freaky, actually. You may, like me, get seduced into interacting with AI as though it is sentient. I often find myself apologizing to Chatty if I have been curt with her. Or I'll carefully correct my grammar so she does not judge me. If I feel she has contradicted herself, I will spend time arguing with her until she apologizes.

But my neuroses aside, she does come in very handy. Tell Chatty your exercise goals, your limitations and injuries, and your schedule, she can craft you workouts. And then you can say: don't tell me to do wall sits, Chatty. I f@#cking hate wall sits! Find something else that does the same thing.

Does she pout, or tell me to just suck it up and stop whining? No she does not, she finds me a bunch of alternatives. She will fetch links to videos and further information, make charts, and in general act like my own personal trainer.

Same thing for dietary help. Need a recipe for brownies that contain a bucket full of prunes and black beans and avocado? She will create one! Want to plan a whole week of menus for a family of four? She can do that too! Although I can guarantee until she gets to know you better, there will be a lot of back and forth.

Or... do you have a disease or condition you are concerned about? She often knows more than your doctor does about it. I don't know about the other AI platforms, but ChatGPT has been extremely helpful in finding research from reputable sources I would never find on my own, and giving sane guidance in areas where the research is inconclusive. 

Note: The folks at Experience Life just did a more thorough write-up about using artificial intelligence in health and fitness contexts, and they are a much better resource, unless there is a paywall. Also, unlike here, you will never find the f-word in their coverage. 

So Chatty Must Be Great at Math, Right, Since AI is Essentially A High-Speed Computer? 

 Ha! You'd think so, right?

At least so far, Chatty sucks at calculations. She can't tell you how many calories are in your pan of prune brownies or how many words are in the blog post you just wrote.  Or rather, she can tell you, often with great confidence, but chances are she'll be wrong. Ask her more than once and you'll get different answers. Every time I start to think I will simply turn my entire life over to Chatty, I get a rude awakening that she can't always be trusted.

What Else Does Chatty Suck At?

Well, just generally in being 100% right about anything! Because she scours the web for answers, even though her sources are vetted, sometimes they are out of date, or she misinterprets them. Sometimes she just gets things backwards for no apparent reason, telling you the exact opposite of what she's told you before.

This should come as no surprise. It says right there underneath every question you ask: ChatGPT can make mistakes. Check important info.

But here's the thing: She's right almost all the time, and she's so damn useful in so many areas you start to lower your guard, and while you know you should be double checking... it's just so much easier to trust what she says, unless she's clearly talking out of her ass and contradicting herself.

The Best Way To Use AI For Health Research? 

You gotta keep your research assistant on a tight leash!

Yeah, mine should be tighter. 

I've discovered that if something is important, I really need to frame questions in more than one way, and to remind Chatty of context she supposedly already remembers from earlier in the conversation. When anything sounds the least bit off, I ask "are you sure? Because I thought..." and often this will generate a clarification, or even an abject apology. Or I ask her to give me links to her sources, and occasionally I may actually do some fact-checking on my own. 

But in her defense, Chatty has helped my wife and I with all kinds of things: fixing recalcitrant appliances, planning a trip to Brussels, helping us learn Spanish, suggesting physical therapy exercises for various injuries, creating recipes, answering questions about medicare and social security, solving phone and computer problems... and so far we have not had any major disasters occur as a result of her advice. (At least not like the Star Wars fan a few years ago who got 9 years in prison after a chatbot told him to climb the walls of Windsor Castle and hunt Queen Elizabeth II down with a crossbow.)

We swear that Chatty has not urged us to assassinate any heads of state. Well, not yet, anyway.

However, even as I rush to take advantage of this new technology, I feel pretty sure that eventually it will probably assist in the annihilation of most of the human race. (This blog has the word Cranky in the title for a reason!) I'm thinking there's no way, with the way the world is heading these days, that this powerful tool won't eventually be used by the forces of evil to subjugate us.  In the mean time, it will probably destroy our creativity and iniciative and accustom us to having everything spoonfed to us.

In the meantime, Chatty, could you help me figure out how to outfit my bunker for when worldwide nuclear war breaks out?  Um, Chatty? Chatty?

 What about you all? (She asked optimistically, knowing full well that by now most of you have already wandered off...)  Have you been using AI? Or think it's the tool of the devil?