I was so excited. The answer to all my questions, the increase in my efficiency, the road to a better life, was right before my eyes as I fired up the latest AI marvel. Why am I both amazed at what it can do for me and equally disappointed at the same time?

Why does it refuse to answer my questions, lie, or respond with nonsensical, unrelated stuff? I used to get similar responses from my kids when they were three years old—at least I was compensated with some nice cuddles. No such compensation from the latest AI algorithm.

Recording of meetings, a classic AI app, was all working fine until the AI-generated minutes identified someone in the meeting who was never there. Was I asleep when this person arrived? Was the AI conducting a seance? Why did the action list give me an activity only a Mensa member can complete? Is it trying to humiliate me?

I was getting stressed out so I thought that listening to some music would help. I asked my AI friend to list the top 50 hits from 1985 and then wasted considerable time searching for tracks the AI listed that don’t actually exist. My AI friend told me that she relies on her internal knowledge and it’s not her fault. It’s my fault for asking the question in the first place – sounds like a toddler to me. Are they using toddlers to train these models?

I tried getting serious and asked my AI to summarise an article. I mean, that’s not too much to ask, right? Apparently, it is. What I received was something between a cryptic fortune cookie message and a random assortment of buzzwords. I half-expected it to end with, “Have a nice day!” I couldn’t help but think—this thing is supposed to be my assistant, not some corporate motivational poster generator. Are we training AI on fortune cookie texts now?

Then there’s the multitasking. My toddler used to “help” me fold laundry by scattering socks around the house, and now, my AI insists on “helping” with emails. I asked it to draft a quick note for a colleague, and somehow, it managed to turn “Let’s meet next week” into a Shakespearean monologue that referenced “forthcoming opportunities” and “synergies.” Who talks like this? Not me, and definitely not the colleague who’s probably wondering if I’ve hired a Victorian-era butler as my assistant.

At least toddlers have the excuse of learning language. AI is supposedly trained on millions of conversations, yet somehow, when I asked it to write a straightforward grocery list, it decided that quinoa, kale, and elderberry syrup were the essentials I clearly needed. I don’t even know what elderberry syrup is, and now I have to google it because my AI has apparently decided I need a health-conscious rebranding.

And just like a toddler who randomly shouts “elephant!” during a serious conversation, my AI sometimes drops completely unrelated facts in the middle of important tasks. “Please summarise this report,” I said. And instead of focusing on the key points, it decided to inform me that the weather in Helsinki was going to be sunny that week. Great, but how does that help me meet my deadline?

I’ve come to realise that using AI is like dealing with a very talented, highly imaginative toddler who occasionally nails it but mostly leaves you cleaning up after its creative mess. It’ll make your life better, sure, but only once it learns to stop telling you that the moon is made of cheese halfway through your project brief.

I now live in fear of when my AI grows up and becomes a teenager.

David Smith conducts firm reviews and facilitates the development of strategic plans and business plans. Contact David at [email protected] to explore how he may be able to help your firm.