Which LinkedIn profile picture is the best?
I updated my LinkedIn profile photo recently. I rarely change it, and I don't have any good criteria for which picture to use besides my personal vibes. I don't want to bother real people for their opinions, so I got some LLMs to diversify my vibes.
Which of these photos is the best profile photo for my LinkedIn profile? I am a Principal Product Manager, primarily for cybersecurity, in the tech industry.
For each model, I used the latest, largest version (ChatGPT 5.2 Thinking, Gemini 3 Pro, Claude 4.5 Opus) with Thinking Modes turned on. Then I threw all 9 responses into Claude and told it give me a markdown table summary.
| Photo | Times Ranked 1st | Times Ranked 2nd | Consensus |
|---|---|---|---|
| Top Left | 9/9 | 0/9 | Unanimous winner |
| Bottom Left | 0/9 | 6/9 | Runner-up (ChatGPT/Gemini) |
| Top Right | 0/9 | 3/9 | Runner-up (Claude only) |
| Bottom Right | 0/9 | 0/9 | Consistently last |
So of course, I went with the unanimous winner. Since my active profile picture was the worst of the bunch, I'm glad I'm updating it now while I'm on the market.
And before you ask: yes, I used AI to enhance my current profile image. Here's the unadulterated image (well, as unadulterated as photos go these days).
The AI photo takes some liberties with my skin tone, lighting, and the background has been smoothed out. It's still me, and is a real representation of what I look like. Would you believe that I was wearing makeup in my original photo? It was a professional photoshoot with a backdrop, supplemental lighting, cosmetics experts, the whole works. I feel that level of presentation is on par with using an AI-enhanced image for your profile picture.
Is this outcome better than a professional photoshoot? It's certainly cheaper and takes less time, but I think the AI photo wouldn't be as good if I were take this same photo with professional photographers today. However, that's subjective and there's no real way to gain consensus on it. Unless you asked each LLM three times, maybe.