Yes, LinkedIn runs a detector on profile photo uploads, but it hunts for one thing: faces of people who don't exist. LinkedIn announced in October 2022 that it checks new profile photos for signs of AI generation, and its 2023 study with UC Berkeley's Hany Farid caught 99.6% of StyleGAN-generated faces. What its rules ban is a photo that isn't you, not a photo made with AI. An AI headshot generated from your own selfies, like Narkis.ai ($27 for up to 200 photos), shows your real face, which is exactly what LinkedIn asks for.
What LinkedIn's rules actually say
- Profile photo guidelines: "You can use an illustration, caricature, or other artistic rendering of yourself, but your profile photo must reflect your likeness." LinkedIn removes photos of logos, landscapes, animals, celebrities, stock images and "others' likenesses". AI isn't on the list. (LinkedIn Help: profile photo guidelines)
- Professional Community Policies: no fake profiles, and "do not use an image of someone else, or any other image that is not your likeness, for your profile photo." (Professional Community Policies)
So the test is likeness. A photo that looks like you passes, however it was made. A face that looks like nobody, or like someone else, doesn't.
What LinkedIn's AI detection looks for
- Fake-account faces. The detector exists to stop fake accounts that use a freshly generated face of a person who doesn't exist, because those faces beat reverse-image search. The 2023 LinkedIn and UC Berkeley paper flagged 99.6% of StyleGAN faces while wrongly flagging 1% of real profile photos. (James Verbus, LinkedIn Trust Data team)
- Not your AI headshot. The same paper reports the method failed on Stable Diffusion faces, and the follow-up study ("Finding AI-Generated Faces in the Wild") notes its model "is not used to take action on any LinkedIn members". (arXiv 2311.08577)
- The "Cr" label is for posts. Since May 2024 LinkedIn shows a small "Cr" Content Credentials icon on images in posts that carry C2PA metadata saying how they were made. It reads metadata a tool attached; it doesn't scan your face. (Social Media Today)
Can people tell your LinkedIn photo is AI?
The real check isn't LinkedIn's software, it's the recruiter who meets you on a video call. If the photo shows a different jaw, skin or age than the person on camera, that is what reads as fake. Try it yourself: our AI or real blind test mixes real photos with AI ones and shows how often people guess right.
How to use AI photos on LinkedIn safely
- Generate from your own face. Upload 15 to 20 recent selfies; Narkis trains a personal model on them in a median 6.5 minutes, so every photo is you in new lighting, clothes and backgrounds.
- Pick the one that looks like you today. Same hair, same build, same age. Drop anything that smoothed you into someone else.
- Keep it plain. Head and shoulders, face filling most of the frame, clean background. Check it with our free LinkedIn photo checker and size it with the LinkedIn photo resizer.
- Don't use a one-photo "glow-up" filter. Filters that reshape your face are the ones that stop looking like you.
What it costs
One-time packs from apps that train on your face, read from each vendor's pricing page in November 2026:
- Narkis.ai: $27 for up to 200 photos ($0.135 a photo), LinkedIn and dating photos from the same model
- Secta Labs: $49 for 200+ photos
- Aragon AI: $35 for 40 photos
- BetterPic: $35 for 20 photos
More: are AI headshots acceptable on LinkedIn, what recruiters think and Aragon alternatives.
Verdict
LinkedIn detects AI faces of people who don't exist. It doesn't ban AI headshots of you. Use photos generated from your own selfies, keep the ones that look like you, and you're inside LinkedIn's likeness rule. Get 200 AI headshots for $27.
