Your Best Models Are Your Staff: Ditch the Fake People Photos

Open your association’s website and look at the people. The smiling professionals in the hero image — have any of them ever attended your conference? Sat on your board? Paid dues? If your site runs on stock photography, the honest answer is no. Those people don’t exist in your world. They’re models, hired for an afternoon, rented to a thousand websites.

Now there’s a cheaper version of the same habit: AI-generated faces. Type a prompt, get a “member” who never lived. It costs nothing, takes seconds, and it’s the most expensive imagery choice an association can make — because associations run on trust, and nothing spends trust faster than fake people.

Real faces win on trust. The research says so plainly.

This isn’t a vibe; it’s measured. Summarizing the existing literature, researchers writing in Frontiers in Communication put it directly: authentic photos of real employees foster greater trust than staged or stock images. Photographs are how visitors infer what an organization values and whether they fit. When the faces are rented, the inference is rented too.

Think about what your website is for. A prospective member is deciding whether this group is their people. A sponsor is deciding whether your audience is real. A young professional is deciding whether anyone in the photos looks like them. Every one of those decisions gets easier with real faces and harder with fake ones. Your members joined a community of actual humans. Show them.

The AI version doesn’t just fail — it becomes the story

You might assume AI-generated imagery is a safe shortcut: no model fees, no consent forms, infinite variety. Researchers at the University of East Anglia tested that assumption against reality, analyzing 171 AI-generated images used by 17 charitable organizations — including Amnesty International, Plan International, the WHO, and WWF — and more than 400 public comments on them. The result should end the debate: fewer than one in five comments engaged with the humanitarian cause at all. The rest argued about the AI — 141 comments on ethics and authenticity, 122 critiquing image quality.

Read that again. The organizations bought efficiency and paid with attention: the audience talked about the fakery instead of the mission. As co-author David Girling put it: “The moment when audiences start questioning whether what they are seeing is real, the emotional connection that drives support is put at risk… organisations could lose far more in trust than they gain in efficiency.”

That’s the dynamic in a sentence. AI-generated people don’t illustrate your cause; they become the topic, and the topic is your credibility.

There is a legal line, and fake testimonials cross it

It gets sharper when the fake face comes with a quote. The FTC’s 2023 update to its Endorsement Guides explicitly expanded the definition of endorsements to cover “virtual influencers — that is, computer-generated fictional characters” and “non-existent entities that purport to give endorsements.” The guides are the FTC’s reading of what counts as deceptive under federal law.

So: a testimonial from a “member” who doesn’t exist isn’t just tacky. It’s the exact scenario the regulator wrote the rule about. If your site wouldn’t put a fabricated quote next to a stock photo model’s face, it shouldn’t do it next to a generated one either.

The uncomfortable nuance: the fakes test better

Honesty requires the other half of the finding. In a well-known 2022 study, Lancaster and UC Berkeley researchers found that people shown AI-generated faces classified them correctly only 48.2% of the time — worse than chance — and rated the synthetic faces 7.7% more trustworthy than real ones. The authors’ own summary: synthesis engines have “passed through the uncanny valley.”

So the fakes can outperform real photos in a lab. Here’s why that doesn’t matter: the trust is borrowed, and it has an expiry date — the moment of discovery. The UEA study is what happens after the lab, in public, when people find out. Trust you have to hide isn’t trust. It’s a liability with a timer.

Besides, there’s a name for what audiences feel when they clock it. It’s called slop — the cheap, synthetic filler now flooding every feed — and people cringe because they can feel the contempt in it: we didn’t think you were worth the real thing. Your members are worth the real thing.

Your best models are already on payroll

Here’s the good news: the fix isn’t a purchase. It’s a camera pointed at people who exist.

Your staff are your best models. Your members are your second-best. Real photos do double duty that stock never can: they show prospective members their future peers, they make current members feel seen (being featured is a quiet member benefit), and they make your organization look like what it is — a living community, not a template.

“We have no photos” is a temporary problem with a permanent solution: the next event. Conferences, chapter meetings, volunteer days, award ceremonies — these are studios that assemble themselves. Ten real photos beat a hundred stock ones, because ten real ones are true.

In the interim, if you truly have nothing: use no people at all. A venue shot, a detail of the work, an empty stage before the keynote — honest and atmospheric — beats a fake person every time.

The reason associations default to stock usually isn’t philosophy. It’s logistics: hundreds of event photos, nobody to sort them, and the vague fear of photographing the wrong person. All solvable.

Consent first. Borrow the checklist member organizations already use: mention photography in invitations and promotional materials, post signs at entrances, announce it before the event begins, and set aside a seating area that stays off camera for anyone who’d rather not be seen. For events where attendees have an expectation of privacy, get written permission from each person in advance — and if someone later asks for a photo to come down, honor it. Consent isn’t bureaucracy; it’s the same trust you’re trying to build, applied to the process.

Shoot real moments. Phone cameras are enough. Candid beats posed — people talking, laughing, mid-applause. Ask a volunteer with a good eye to be the event photographer; it’s a role people enjoy.

Cull fast. This is the actual bottleneck: 800 photos from the annual meeting, and someone has to find the 40 keepers. Our founder, Joseph Watkins — an avid wildlife photographer — built CullPro.ai for exactly this problem: it scores every frame on your own machine, explains each call, and hands you only your strongest work. Free forever, nothing uploads. Camera to canvas in minutes instead of a weekend.

Publish and rotate. Swap the homepage hero seasonally. Feature a member a month. Retire anything that starts to feel like wallpaper. A site whose faces change is a site that’s alive.

Show the world who you are

Every association is, at bottom, a group of people who chose each other. Stock photos say we couldn’t be bothered to show you us. AI faces say we invented someone instead. Real photos say the thing that’s actually true: here we are.

That authenticity is a strategy, not just an aesthetic — we’ve written before about what training is actually worth paying for, and the throughline is the same: the things that build trust are the things that are real. Practice over certificates. Members over models. Show the world who you are, and let the world decide it wants to join.