Methods for Using AI

Brief notes from a sceptic, and the least techy person you know.

A beginner’s tale

I used AI for the first time a few weeks ago.

Wait, is that accurate to say? In case it matters, I used ChatGPT-4 for the first time.

I searched it on Google. Logged in via my Google account. And bam. A chat screen where I could ask anything…

But I didn't actually know what I wanted to ask it. I had landed there on account of an encouraging colleague who, when I'd asked for advice on AI and all I needed to know about integrating it into a new project I am all of a sudden working on, suggested the radical act of actually trying it out myself.

I have resisted any interaction with AI for so long. Several people in conversations about this (Top Secret!) aforementioned work project have suggested that the Paper Clip from Microsoft Word (Clippy?) in days gone by was an early form of AI. Again, I don't know if that's true, and Google isn’t giving me a straight answer. AI or not, I did of course have many an interaction with that friendly bit of stationary.

But when it comes to all this OpenAI, ChatGPT, Claude etc etc., I know nothing. I don't know what's a brand, what's software, what's a tool that many companies are using verses what is owned by one. I've not engaged in the hourly news updates around it, or the how-to guides on perfect prompting, or fanfare reports of all the wonders it can do. I thought I might leave it all blow over, until no one's bothered by it anymore, and I'll have not had to expend any brain power on learning what it is or why it is important. But, both because of my work, and, ahem, general world vibes, it's starting to feel like that logic will turn out much the same way as my approach to COVID in early March 2020. I suggested we had bigger things to think about and there was no need for people to keep banging on about it. A suggestion that, of course, became entirely defunct as it took over the whole world…

Generating schedules

So there I was a few weeks back, logged into ChatGPT-4, not knowing what to expect. There were a few prompts lined up for what I might want to ask. I went for the one that felt most closely aligned with what I needed to know in that moment.

I clicked on "Plan a relaxing day."

The AI duly leapt into action, mapping out a luxurious day of activities like yoga and nature walks, midday meditation and massages, with fully scheduled time slots and brief descriptions of each activity, should I need to get a bit more of an idea. All this unfolded in front of my eyes before I would have had the chance in real time to make a list of the kind of activities I might find relaxing. ChatGPT did the thinking for me.

Inspired by the curated timetable, I then went in with my own question. Could it help me plan out my own writing retreat? Enthusiastically it responded,

"Absolutely! A writing retreat can be a wonderful way to immerse yourself in your creative process. Here’s a plan for a writing retreat that combines focused writing sessions with relaxation and inspiration activities."

before gushing out not only a daily schedule, complete with breaks, snacks, and nourishing activities, but also tips on how to prepare (think location and goals) and additional notes on staying hydrated, and rewarding myself for small milestones.

Now, it's not that I think the AI got it spot on. Even though it clearly has a good sense of what a writing retreat is, and indeed understands the repertoire of classic activities involved, it doesn't know me. It doesn't know what I like or enjoy, or that I don't ever eat breakfast as soon as I wake up, much preferring a brunch slot. Or, when deep in writing, the biggest almond croissant you can find that I can nibble on over the course of a few hours while drinking copious amounts of coffee.

But it did give me a pretty instantaneous structure, which I wonder perhaps has merit for saving my brain some time and energy. Rather than expending it on the planning, which someone (or something?) can do for me, I'm free to concentrate on the task at hand - whether it be relaxing or writing or whatever else I might think I need a schedule for.

A desperate plea

A few days later, in a moment of stress and anxiety at everything on my to-do list, I was back again. I wanted it to sort my life out for me. Pronto.

The prompt I used was “can you help me create a daily schedule for getting done everything I want to do, and balance that with my energy levels?” Of course it could. It asked me a few brief questions – this time to get to know me more. What I need to fit into my day. How my energy levels move at different points in the day. Do I have any time constraints. In my imagination, I liked the thought of someone on the other side of the screen, clipboard and pen at hand, poised and ready to act as a helpful assistant to my dilemma / momentary freak out about getting it all done.

Once it had my answers, within a minute it shot out a daily schedule, encompassing everything I told it I needed, broken down into slots of 30-60 minutes. I was impressed, but as I dug into the detail, it wasn’t exactly spot on.

It had managed to chop an hour off of my required working hours (not sure the boss would be too happy about that) and again hadn’t taken into account my eating habits (breakfast at 6:30am? No thank you). But it did offer up some suggestions I’d never have considered, like 30 minutes of exercise at 3:30-4:00pm, when I’m often in the depths of the afternoon slump, struggling with concentration but determined to be productive. Sometimes I’ll be in a meeting, or on a COP deadline, that would make it near-impossible to just hop off for a bit of exercise, let alone find the motivation (can AI generate that?!). But maybe once or twice a week it is something I could consider, getting a bit of movement in then to keep me going for the rest of the day, even if that means working 30 minutes later to make up the time of course. Or perhaps taking a shorter lunch break, to fit in some mid-afternoon movement.

Admittedly I’ve not yet fulfilled my intention to trial at least one day that follows my AI-generated schedule, but it has at least got me thinking.

And maybe that’s what AI is all about.

Getting the cogs turning

In my day job, I’m being encouraged to think outside the box with AI. Get creative. Think about how it can help us solve challenges, speed up processes and provide richer experiences.

I’m working with developers whose brains baffle me on a daily basis, and whose enthusiasm for the opportunities AI brings is rather infectious. Elsewhere, others are thinking about how it might revolutionise our entire sector of education, and, indeed, whatever sector you may work in too.

I’ve recently started reading

Ethan Mollick

’s

One Useful Thing

, which offers some digestible snippets on this stuff.

Dr Nia D Thomas

shared an interesting thought piece this week on balancing humanism with ultra-digitalism on

Knowing Self Knowing Others

. Graham Allcott writes a bit about all this too, and has an AI powered chat bot, compiling the work-culture knowledge he has previously shared through writing, which I’ve been known to dabble with. Helen Tupper and Sarah Ellis, the duo behind Squiggly Careers, did an interesting podcast episode on using AI for career development. There are many others doing wondrous things for the contemplation of AI I’m sure, and if you have any recommendations for thinkers to read, or listen to, do share in the comments.

I’m grateful for these thinkers, these writers, for doing this work. Because I, frankly, can’t be bothered. Yes I’ve watched Sal Khan’s Ted Talk on personal tutors and chatted with colleagues in amazement. But thinking about the implications of that potential over the next five/ten/fifteen years kind of makes my brain hurt, and if I’m honest, invokes a bit of fear in me.

I mean, I’m really not the target audience for innovation. I still carry cash, am sceptical about having my bank card linked to my phone, and never will I ever own a Kindle. If I need customer support, I want to call and speak to a human. I don’t like online shopping, even if it means saving time and/or money. And wifi on planes? Please. Do not get me started.

So, I’m learning to wear a different hat at work, pushing myself to be more adventurous with AI in the context that I’m needing to use it there, while in my personal life maintaining a healthy dose of avoidance and analogue-ism. That said, it does seem to be slowly seeping in. Last week when discussing an issue with my partner, I suggested we present the problem to ChatGPT and see what it had to say. As a tech enthusiastic he was as shocked as he was delighted at my out of character proposal. As an old-school lover of discussion and jotter-of-notes, I paused and wondered who I was becoming…

Playing around

My partner had suggested recently I might look to use AI to generate images for my Substack, offering a stylistic consistency. I get the point, I like the aesthetics of others who take this approach, whether generated by artificial or human intelligence, but it’s not for me. I like sharing photos here, real life snapshots, from, well, real life. And if I were to go down the illustration route, I’d personally rather work with a designer or illustrator on it.

That said, in the spirit of play and experimentation, as I worked on this essay, I did dabble with AI image generation on Canva, and after a few rounds of prompt tweaking, came out with this take of me in the lemon dress I wrote about recently, enjoying a coffee … in a lemon grove apparently. Despite specifically saying in the prompt “no lemons in the foreground or background”, I could not get past the AI’s enthusiasm for matching the dress with the surroundings.

When life gives you lemons …

I’m sure there are ways around it, but I just don’t have the inclination to keep trying. It’s annoying, rather than intriguing. I don’t want to have to crack the code. I want it to just understand me. Like a human would.

That said, if you are into code cracking and wanting to learn more about this kind of image generation,

Marie Vandoorne

is doing some cool stuff over on

The Bored Millennial

.

Elsewhere,

Beth Spencer

at

Introvert Drawing Club

, who I recently enjoyed listening to on

Evelyn Skye

’s podcast, has created a badge, and really a movement, for championing human intelligence in art.

And isn’t that the beauty of Substack? Something for everyone to explore and resonate with.

Moving forward with AI/HI

Returning to where I am with AI without my work hat on, instead perhaps wearing my writer / normal person trying to get through the to-do list hat, I don’t see myself integrating AI into my life in any significant way – and certainly never related to writing.

As an obvious technophobe, the idea that the words here on

The Navigation

would ever be anything other than written with human intelligence is truly laughable.

But, maybe I can be open to the fact that it could save me time. That AI could save me a bit of brain power that I can put to better use, either for my writing here, or being creative in other ways. The writing retreat structure that ChatGPT provided is still in my mind to try out, and while I could design a more personalised day myself, at least this offers a starting point to take what works and leave the rest. Similarly with the suggestions for how I might structure a typical workday. Sometime soon, maybe even this week, I’ll make a point of trying out the suggested schedule, complete with a little afternoon workout, and see what happens. Will it revolutionise my flow? Unlikely perhaps, but if it helps a little, and provides some form of energy boost, where’s the harm.

Now, of course, neither of these examples need come from AI. There are undoubtedly countless articles online and in print, and probably even here on Substack, written by humans, sharing ideas for such schedules – and I am not suggesting we shouldn’t read them. I for one love that sort of writing, learning what works for others, and getting ideas for ourselves.  

Which is all I’m trying to do here really - if the notion of a quick fix AI search/idea generation session speaks to you, give it a go. And if not, leave it be.

In any case, thank you for making it this far with the sceptical ramblings on the methods and approaches I’ve been taking as a self-confessed technophobe.

Viva la revolucAIón? Let’s see.

If you liked this piece and would like to show some support without a paid subscription, a Ko-fi coffee is always appreciated.☕

Thanks once again to

Claire Venus

and all at the

Sparkle on Substack

24 Essays Club for the encouragement to experiment, write, play.