Job application check

How to spot a fake job posting with AI photos

A fake job posting works on trust: a convincing advert with a polished website, professional team photos and a clean description. AI makes those parts cheap and fast — generated office scenes, invented colleagues and fluent text that cannot be verified against anything. Photos and text therefore prove less and less on their own. Always verify the company, the channel and the conversation outside the advert.

Direct answer: five checks on a job posting

1. Verify the company outside the advert. Find the organisation through its own website or LinkedIn, not through the link in the advert. A real company you recognise can be impersonated: check the email domain and phone number on the official site instead of using the details in the advert.
2. Reverse-search the photos. Run the team or office photos through reverse image search. The same image under another company name or from a stock service is a strong signal. No match proves nothing, however: brand-new AI images have no findable source either.
3. Check the photos for AI signals. Look at faces, fingers, text on signs and walls, reflections and repeated patterns. Several details that shift between images justify follow-up. Use the image detector as triage: a strong warning means extra checks, a quiet result proves no authenticity.
4. Check the channel and contact details. A posting you only know from an unexpected message, a private app or an unknown job board deserves extra checks. Be especially alert to requests to pay, to cover administrative fees or to sign a contract outside the company.
5. Ask for a real conversation. A genuine employer can arrange a video call or a phone call through a number you found yourself. A fabricated company often offers chat or email only and avoids a direct conversation.

The US Federal Trade Commission has long warned about job offers that lead to identity theft or money loss, with signals such as unexpected messages through chat apps and payment requests before a contract (FTC, job scams). The pattern is global, and AI-generated photos and text make it cheaper to stage.

Why AI makes this harder

Poor translation and amateur layout used to be warning signs. A language model now produces a flawless posting, and an image generator makes consistent company photos. The check therefore moves from presentation to the reality behind it: does the company exist, does this role exist, does it use this domain and these channels, and can you speak to an actual person?

Use the detectors as triage

Upload the photos to the image detector and paste the posting text into the text detector. Both give a model-bound warning signal only: a high score proves no fraud, and a low score proves no authenticity. A generated photo can appear in a real posting, and a real photo in a fake one. Always combine the signal with the checks above.

Decide from the follow-up

If you reached the company through a number or domain you found yourself, the photos are traceable and nothing is requested in exchange before a contract, you have several independent signals. If it stays vague, a conversation is avoided or money is requested, report the posting to the platform and share no further data. Employers reviewing incoming applications can use the recruitment review workflow; for imagery the general guide on telling if an image is AI-generated applies.

Frequently asked questions

Does a high detector score prove a job posting is fake?

No. The score is a model-bound warning signal and says nothing about whether the company or role exists or about the advertiser's intent.

May I upload photos from a job posting?

Public posting photos are usually fine as long as you do not redistribute them and they contain no other people's personal data. Do not share a detector score as an accusation.

What is the strongest sign of a fake posting?

A payment request before a contract, contact that only works through chat or a private app, and a company you cannot verify through official channels.