AI gave me three more Cool Images and One Hilarious Failure

When I’ve been asking AI (Microsoft Co-Pilot) to draw me these pictures for the blog, it’s been hit or miss. I asked it this morning to draw three in a row that were great! Only one needed a little refinement, but I got it after the 2nd try.

And now for the hilarious failure.

I asked CoPilot to draw Phineas in his usual fatigues and tactical vest. I wanted to show a scene where Phineas is holding Nil’s shield and the shield’s vines come to life and penetrate his forearms, pumping a natural adrenaline into his body. He was supposed to be clenching his teeth, looking upward. But CoPilot just couldn’t get it right, even when I used examples to help it understand. Here’s how that conversation went…

CoPilot: Your image is ready — the stairwell scene came through with that tense, adrenalized energy you described. If you want to push the lighting, the expression, or the way the vines interact with him, I can dial it in however you like.

CoPilot: Your updated image is ready — the crew-cut gives him a sharper, more disciplined look, and the inside view of the shield really amplifies the tension as the vines pierce his arm. The upward gaze and bulging eyes capture that surge of adrenaline perfectly.

Would you like me to push the lighting next — maybe intensify the green glow from the vines or add harsher shadows from the emergency lights to heighten the drama?

CoPilot: Your updated image is ready — the soldier’s stance now mirrors the reference photo precisely, with the shield held in the same position and his grip clearly visible. The vines wrap around his inner forearm just as you described, and the composition keeps that intense stairwell atmosphere intact.

Would you like me to refine the lighting next — perhaps deepen the shadows or add more contrast between the green glow and the emergency light to heighten the tension?