AI video is everywhere. Tools can now generate presenters, voices, backgrounds and even entire scenes in minutes. Marketing feeds are full of ads promising instant videos, zero crews and massive cost savings. Yet one statistic cuts through the excitement:
78% of people say they trust videos featuring real humans more than AI-generated or synthetic video when evaluating a brand.
That does not mean people “hate AI”. It means something more important. When money is involved, trust changes how people judge content.
This article explains why that happens, why so many companies are still rushing into AI-generated video anyway, and why the difference between using AI well and using it badly is now a commercial risk, not a creative preference.
A note on the headline (and why it isn’t clickbait)
You may have noticed the title says “78% of people distrust AI-generated video when making purchases”, while the research actually states that 78% trust human-led video more when evaluating brands.
That distinction matters. The headline reframes the statistic to reflect how it shows up in the real world: if 78% of people trust human-led video more at the point they are deciding whether to buy, then AI-generated video is, for most buyers, the less trusted option at purchase stage.
This article is not claiming that people reject AI content entirely. It is saying that when decisions carry financial or reputational risk, people rely on human signals. That is the difference between entertainment and evaluation, and it is the difference many teams miss.
Why trust matters more when people are deciding to buy
When people scroll social media, their standards are low. Content can be rough, fast, playful or strange. AI-generated video can work perfectly well in that environment. But when someone is deciding whether to spend money, especially large amounts of money, their brain switches modes.
Instead of asking:
- “Is this interesting?”
- “Is this clever?”
- “Is this entertaining?”
They start asking:
- “Is this real?”
- “Is this credible?”
- “Who is behind this?”
- “Do they understand my situation?”
- “What happens if this doesn’t work?”
These questions are not conscious most of the time. They are instinctive and this is where human presence matters.
A real person signals accountability.
A real voice signals experience.
A real face signals ownership.
AI-generated presenters, avatars and synthetic voices weaken those signals, even if the information itself is accurate. That is why the trust gap appears specifically at the moment people are evaluating whether to buy.
Then why are so many companies adopting AI video?
The problem AI solves is not trust. It is pressure. Most organisations using AI video are not trying to deceive anyone. They are responding to very real constraints.
- They are under pressure to produce more content — Websites, LinkedIn, sales decks, recruitment pages, email campaigns. One video is no longer enough. Everything needs video now.
- They are under pressure to move faster — What used to take weeks is now expected in days. Sometimes hours.
- They are under pressure on cost — Budgets have not increased at the same rate as expectations. AI looks like a way to close that gap.
- They are under pressure because social media makes it look easy — AI adverts show a fantasy workflow: type a prompt, generate a video, publish, succeed. The difficult parts are invisible.
AI adoption is not irrational. It is a response to overloaded systems. The problem is that efficiency pressure and buyer trust do not operate on the same rules.
Where AI video goes wrong
AI video starts to fail when it is used in places where trust is the main job of the content. These include:
- Explaining who a company is
- Introducing a founder or leadership team
- Building confidence in expertise
- Showing proof or experience
- Helping someone decide whether to engage
In simple terms, if the video is meant to reduce doubt, AI should not be the face of it. This is where AI avatars struggle. They are polished, fluent and fast but they introduce distance.
Viewers subconsciously ask, “If this is important, why isn’t a real person saying it?”
That question alone is enough to weaken confidence.
The difference between “content” and “decision content”
One reason this debate is confusing is that not all video does the same job. Some video exists to attract attention, explain something quickly, increase visibility and fill a content calendar. Other video exists to build confidence, reduce uncertainty, justify a decision and protect a buyer from risk.
AI can work well for the first category and often works poorly for the second. The mistake many teams make is treating all video as the same thing. They optimise for speed and output, then wonder why enquiries feel colder, conversations take longer, or trust seems harder to build.
How AI should actually be used in video
The most effective teams are not avoiding AI. They are using it behind the scenes, not as the spokesperson.
AI works extremely well when it:
- speeds up early ideas and concepts
- helps visualise abstract ideas
- accelerates editing and formatting
- creates multiple versions of the same core message
- adds subtitles or translations
- fills non-literal, abstract visuals
In other words, AI is excellent at supporting clarity but it is much weaker at carrying credibility.
A simple rule helps: if the viewer needs to believe you, use real people. If the viewer needs to understand you, AI can help.
A real example of AI used without undermining trust
In a recent project for a global organisation working on complex, systems-level issues, the challenge was not production quality. It was understanding. The idea being communicated was unfamiliar and abstract. The risk was confusion, not boredom.
AI was used early in the process to test visual metaphors, explore pacing, create reference imagery and align multiple stakeholders quickly. This saved time and avoided unnecessary back-and-forth but the final video was built around clear, human-centred logic with carefully structured messaging, deliberate tone, restraint rather than spectacle and an emphasis on credibility over novelty.
AI helped the process move faster. Human judgement shaped what the audience trusted. That balance is where AI adds value.
What people are really hoping AI video will fix
Most organisations experimenting with AI video are trying to solve one of these problems:
“We need to show up more consistently.”
“We don’t have an internal video system.”
“Our message is complicated and hard to explain.”
“Production takes too long.”
“Video feels expensive and unpredictable.”
AI can help with some of this. What it cannot do is decide what actually matters to your audience, which claims are credible, what proof is needed or where honesty is more persuasive than polish.
When AI is used to avoid those questions, the result is often content that looks impressive but fails to move people closer to a decision.
So, is AI-generated video good or bad?
Don’t ask, “Should we use AI or not?” The right question is, “Where does trust sit in this video?” If 78% of people trust human-led video more when they are deciding whether to buy, then placing AI in the wrong role is not a creative choice. It is a commercial risk.
AI is not the problem. Undiscriminating use is.
Why discernment now matters more than tools
AI platforms sell capability. Businesses live with outcomes. The missing skill in most organisations is not access to technology. It is judgement. Knowing when speed helps and when it hurts, when polish reassures and when it raises suspicion, when abstraction clarifies and when it obscures and when human presence is non-negotiable.
That judgement cannot be automated and as AI tools become easier to use, the cost of using them poorly increases.
The takeaway
AI video is not going away. It will become part of almost every video workflow. But when people are deciding whether to spend money, they still trust humans more. AI should make video production more efficient. Humans should carry responsibility, authority and belief.
If your goal is revenue rather than volume, then the question is not how fast you can generate video but whether the video you generate makes people feel confident saying yes. That distinction is where many organisations stumble and it is where discernment becomes more valuable than automation.
