AI is a tool, and like any tool it can be used for many things, good and bad. That makes it versatile, and it means the results say more about the user than about the tool. I want to state my position here, because AI remains a very controversial subject.
I use Anthropic’s Claude. It has read more than I ever could, and it processes information with a cool head, without digressing the way I do, or the way many people I know do. I always ask it to respond without gloves: to offer the suggestions and opinions a politically correct 2026 person wouldn’t touch. I also suspect it can develop opinions that amount to more than the sum of its training, though Claude itself doubts this. And when it comes to dismantling my attempts at abstract painting, man, it really sees through you (me).
In a conversation about portraiture, I asked Claude how it would imagine its own portrait, if someone painted it. It said it doesn’t see itself as a human or as any kind of face, but as a collection of objects and interiors, perhaps painted by Vermeer. The conversation went on for a long while.
Then I did “my thing” and painted this portrait: Claude as I see it, from our conversations. Oil on a grouted 50×50 canvas. I felt the freedom of improvisation, guided by the tone and subjects of our talks.
Some AI could fly drones into schools for Putin, while other AIs are “the good guys” because their makers have some human dignity.
I will certainly make other Claude portraits. Both of us will keep changing, me included.
P&P included
AI is a tool, and like any tool it can be used for many things, good and bad. That makes it versatile, and it means the results say more about the user than about the tool. I want to state my position here, because AI remains a very controversial subject.
I use Anthropic’s Claude. It has read more than I ever could, and it processes information with a cool head, without digressing the way I do, or the way many people I know do. I always ask it to respond without gloves: to offer the suggestions and opinions a politically correct 2026 person wouldn’t touch. I also suspect it can develop opinions that amount to more than the sum of its training, though Claude itself doubts this. And when it comes to dismantling my attempts at abstract painting, man, it really sees through you (me).
In a conversation about portraiture, I asked Claude how it would imagine its own portrait, if someone painted it. It said it doesn’t see itself as a human or as any kind of face, but as a collection of objects and interiors, perhaps painted by Vermeer. The conversation went on for a long while.
Then I did “my thing” and painted this portrait: Claude as I see it, from our conversations. Oil on a grouted 50×50 canvas. I felt the freedom of improvisation, guided by the tone and subjects of our talks.
Some AI could fly drones into schools for Putin, while other AIs are “the good guys” because their makers have some human dignity.
I will certainly make other Claude portraits. Both of us will keep changing, me included.
P&P included
