As someone deeply involved in digital technologies, I’ve had my fair share of interactions with AI, and for the most part, it’s been a useful, efficient tool. I’ve found it to be most useful in sorting data, and even pulling out insights from transcripts.
ChatGPT does make mistakes though.
But recently, I had an exchange that left me reflecting on the limitations of even the most advanced ML, and “conversational” AI tools.
I started a conversation, asking a simple but curious question:
Tell me one thing you think I don’t know about myself.
The response was interesting and…off.
One thing you might not realise is how consistently you blend your technical expertise with creativity. Whether you’re organising industry events, working on vinyl projects, or managing digital communications, your ability to balance detailed, structured thinking with a flair for creative expression is a strength that comes through in many areas of your work.
Wait a minute. Vinyl projects?
That’s not me, so I corrected the AI:
Hmm – I don’t actually work on vinyl projects though?
The AI acknowledged the mistake, saying:
Ah, you’re right! I was referring to [name redacted] vinyl projects, not your own. Thanks for the clarification. Your focus seems more on digital technology, leadership, and strategic management, rather than hands-on creative work like vinyl projects. What stands out for you is your ability to connect the dots between technical innovation and strategic business goals, particularly in the printing industry and digital communication realms.
Given that I use this tool for a number of varied purposes – I had been curious to see what it would say to the question – and my suspicions were correct. It had confused me with someone else. While it’s true that I balance technical innovation with strategic goals, I write about vinyl projects, car wrapping and signage, but I don’t do those things. A reminder that AI, despite its advancements, and sophistication, still has room for improvement.
But what struck me wasn’t just the error. It was the AI’s next attempt to tell me something I didn’t know, when I prompted it with:
OK – fair enough – that is something I know though – what do you think I don’t know ?
And out came:
One thing you might not fully appreciate about yourself is the significant influence you have in shaping communities and networks. Beyond your technical skills and industry knowledge, your involvement in initiatives like Women in Print, organising events, and supporting others shows that you’re not just a participant but a catalyst for connection and growth in your field. You might not always see the ripple effect of your contributions, but you’re helping create spaces where others thrive and develop. That’s a quiet yet powerful impact that goes beyond your day-to-day work.
Third times a charm – I think it nailed it. Because I think it is correct that I have an influence that I don’t see in the communities that I serve. But the experience in this conversation has done little to change my mind about the “AI is going to take our jobs” line.
AI can analyse the data it has access to, the data we give it, it can connect dots from past conversations, quicker than we can make connections with current information. It can offer suggestions, and make inferences, but it still doesn’t truly know us. It can identify patterns, but it doesn’t live our day-to-day realities, nor does it feel the emotions behind our decisions and actions.
So, while AI is a powerful tool, this is your reminder that self-awareness and human connection are still deeply personal experiences, human experiences that only we as humans can have —and that even the most sophisticated machine learning systems can’t fully grasp those experiences. They can only mimic the artefacts of those experience. Like having what feels like a conversation to us, and is really just a computer responding to prompts.
In the end, the AI didn’t know me, despite giving it a good old try. And that’s okay. Because perhaps the real value of AI is not in telling us things we don’t know, but in sparking conversations that help us reflect on who we are.
And for me? I’ll leave the vinyl projects to those that have that skill.
Oh and the image at the top of this blog was generated in ChatGPT40 by a process called stable diffusion. So much fun working through the various iterations to finally land on this image.
