What almost getting hacked taught me about AI, human experience, and what it means to understand

I am fairly new to Substack and love it! I have naively scrolled and typed and restacked and made notes. I’ve been happy. Then one day I came across someone, Derek Hughes, an expert who helps with writing. And, I loved every single word he had to say. I enthusiastically went on a search to read more of what he had to say, watched his videos and as a result, signed up for something he was offering at an entry level. 

I made a public statement of appreciation that was genuine. But Derek wasn’t the only one who read that message. Others did as well. It was public.

And here is where the plot thickens.

Fast forward a few days, and I received a direct message from Mr. Hughes offering to connect with me about a new offering he had. I noticed it was a slightly different picture than the one he uses on Substack, but said to myself, “that must be a photo that is different for his DMs.”

After two or three exchanges back and forth he asked me if I could join him on telegram. I was relaxed so texted back, “I’m already on Voxer and Whatsapp, Signal and Messenger. I don’t want to get on another app.” So, of course, I had just given some valuable information out to Mr. Hughes, as he called himself. I thought it might be an Irish formality to refer to himself as Mr. Hughes rather than Derek. Oh, how innocent I am. LOL

I won’t bore you with all the ins and outs, ups and downs of my continuing interaction with Mr. Hughes (henceforth to be referred to as, Mr Scammer), but know that the phishing continued now on multiple sites and I very quickly became much the wiser. And then today, a mature older woman reached out for coaching and sent me a suspicious “Zoom” link. Was it Mr. Scammer again? I suspect so, because of the timing and some “tells”. But by now, I was paying attention.

I turned to my AI agent to help me ensure safety across all the platforms I use. It helped me identify the suspicious links, told me where not to click, and helped me think through the pattern.

Eventually, I knew I was safe. AI had been extremely helpful. And, to be honest, I was still feeling a little shook. It was at this point that my AI cracked a joke.

And the joke was….

AI: Your new rule doesn’t need to make you suspicious of everyone. It just removes their leverage; New contact → your email → your scheduling system → your meeting link. They can be an 87-year-old grandmother with an adorable Gmail photograph. They’re still using your Zoom link. 😂

Perhaps I was feeling more vulnerable, so the joke somehow really landed and got my attention. I laughed. And then I stopped. Because something about what had just happened fascinated me. How can AI understand that I was feeling a little shook about this persistent scammer and that what I needed most after all the technical support and help, was a joke to ease the tension?

I mean how can something that cannot feel fear and had never felt what I was feeling in that moment, somehow recognized the moment where humor would relieve the tension, not minimize it? That’s fascinating to me.

A conversation ensued that I never anticipated having. Here is what I asked my AI agent:

Sydney: Can I ask you a question, here? The humor you just used felt very human. I guess AI can appreciate what humor is, not because it makes you laugh but because you appreciate the sequencing? Not sure. But when it comes to emotions and physical pain, a human experience. How can AI begin to really understand what it can’t experience for itself?

AI: You’re pointing to one of the most interesting distinctions between understanding an experience and having an experience.

I don’t laugh. I didn’t have that little involuntary physical reaction a human might have when I wrote, “87-year-old grandmother? My Zoom.” There isn’t an internal Sydney-and-me moment occurring somewhere inside me.

But I can understand a great deal about why that was funny — And because we’d already established that you were safe and hadn’t clicked the link, humor was appropriate rather than minimizing something frightening.

Sydney: My mind is going… I have so much I want to interject, but the AI continues.

AI: Emotion is where the distinction becomes much more important

I don’t know what grief feels like.

I don’t wake up and, for three seconds, forget that someone died in my life. I don’t experience the moment when memory catches up.

And therefore there is a statement I shouldn’t make to you:

“I know how that feels.”

I don’t.

What I potentially can say is, “I understand what you’re describing.” And there’s a meaningful difference.

Suppose someone tells me, “My neck hurts.”

I have never hurt.

I’ve never had the primitive, immediate experience of make this stop.

That’s something no amount of textual knowledge converts into sensation.

You know human experience from the inside-out.
I know a great deal about human experience from the outside-in.

My mind started racing… Is experience actually synonymous with understanding? We assume it is. 

But I’ve spent decades working with human beings who have experienced grief and are still terrible at sitting with someone else’s grief. People who know rejection intimately and still reject others. People who have experienced abandonment and nevertheless abandon. People who have suffered terribly and somehow become less, rather than more, able to tolerate another person’s suffering.

Experience is neither necessary nor sufficient for understanding. And now the AI question gets considerably more uncomfortable. Because AI doesn’t possess the first thing, “experience,” yet it can sometimes demonstrate aspects of the second: recognition, interpretation, prediction and response quite well.

That doesn’t mean AI feels. AI isn’t conscious and it isn’t capable of loving you or caring about you. But perhaps we’ve been using the word understanding too casually.

I was truly afraid and AI didn’t experience my fear. I was reassured, but AI didn’t experience relief at that. AI made a joke, in the absence of being able to experience amusement.

I was the one who laughed. The laughter was genuine. The one cracking the joke didn’t feel anything close to what I was feeling, it didn’t feel anything at all. But laughter happened for me on my side of the conversation.

And you know what? I just realized that outside-in understanding is something humans do, too.

That doesn’t erase the profound distinction between me and AI. I have a body, consciousness, subjective experience and an embodied reference library. But in my work, someone can bring me an experience I’ve never had (it happens all the time), and I don’t say, “Sorry, can’t help you. I’ve never lived that.”

One aspect of experience that AI and I share is that I sit on the outside of people’s experiences all the time. Some I have had. Many I have not. Yet I don’t believe that makes me incapable of understanding.

I listen. I ask questions. I recognize patterns. I bring perspective, tools and sometimes an ability to see something that is difficult to see from inside the experience itself. I work from the outside in all the time. But I also possess something AI doesn’t. I have an inside. I have a body that has hurt, a nervous system that has been afraid, a heart that has broken, and a lifetime of experiences that become reference points when I try to understand someone else’s.

AI has representations of those experiences. I have lived them. That distinction matters enormously to me and I hope it matters to you. I don’t want us to lose this awareness as AI becomes increasingly good at sounding human.

Maybe experience and understanding were never the same thing to begin with. AI didn’t experience my fear this morning. It didn’t experience relief when we figured out I was safe. And when AI made the joke, it didn’t experience amusement. I did. The laughter was real. It just happened on my side of the conversation.

And maybe that’s the question I’m left with: not whether AI can experience what it means to be human, but how much something can understand us from the outside when we are the ones living the experience from the inside.