This is the fourth post in Eight PR's Visionary Series - The Business of Tomorrow, featuring perspectives from visionaries across industries and markets. We are grateful to Tony Jones, Managing Director, Adaptive Media Partners, for agreeing to take part and contribute to this series.

There's a lot of noise around generative AI at the moment - especially synthetic images and video.

Some of it is exciting. Some of it is genuinely useful. And, let’s be honest, a lot of it is crap. Or, to use the more polite internet-age word: “slop".

We're all seeing more synthetic content in our feeds: fake people, fake stories, fake authority, fake outrage, fake news, fake hype. The volume is going up, but credibility and trust is going down.

That's the part we need to take seriously.

Because if we use AI badly, we don’t just make poor content; we undermine the relationship between the audience and the message. And the damage can be deep and long-lasting.

For Earth Day, we worked with Harry Chan, Hong Kong’s “Ghost Net Hunter”, on Lamma Island. Harry is a retired businessman, veteran scuba diver, and ocean conservation volunteer who has spent years helping to remove abandoned fishing nets, plastic waste, and marine debris from the sea.

Harry Chan, Hong Kong’s “Ghost Net Hunter”, Credit: Tony Jones

So, this is what we did. We filmed Harry for real. We captured his face, his expressions, his presence, and his natural way of speaking. We recorded his real voice. Captured his passion. We worked with a script based on Harry’s own words, and where the language was refined for clarity or fluency, it was approved by him.

We also filmed the real issue: drone shots of plastic waste on the beach, ghost nets under the water - all of it real. The debris all around Lamma Island was real. None of the ocean waste was generated by AI. None of the damage was invented. We didn't ask an AI model to “imagine a polluted beach”. We went to the beach and filmed what was actually there. And there was a lot of it.

That distinction matters. Because the ethical question around AI video isn’t simply, “Did you use AI?”

The more nuanced question is, “How did you use AI?” Did it amplify something real, with consent and transparency? Or did it manufacture something that only pretends to be real?

In this case, the AI was the megaphone, not the source.

We used real footage of Harry to create a photorealistic avatar. We used Harry’s real voice to create an AI voice clone, so his message could be localised into different languages. We used real footage of real ghost nets, real plastic, and real marine waste to ground the story in the physical world. None of the waste was AI-generated. This was important to Harry and the team.

The technology helped us take one authentic message and adapt it across multiple languages and versions.

But the raw ingredients were real.

That's the model I believe in: authentic capture first, AI amplification second.

Ocean waste on Lamma Island. Credit: Tony Jones

If we were to put an authenticity “nutrition label” on this project, the ingredients would be very clear:

• Real person.

• Real voice.

• Real script.

• Real location.

• Real ocean waste.

• Real consent.

• Real approval.

• AI-assisted localisation and distribution.

Synthetic slop starts with a prompt and works backward toward something that looks plausible. It often has no real person, no real place, no real experience, and no real accountability behind it. It fills a space. It might make audiences lean forward until they realise what they’re looking at. And over time, all of these artistic licenses erode our trust in public relations, that “seeing is believing”. A trust that’s been built over generations of human documentation, news gathering and storytelling.

AI-amplified authenticity starts from a different place.

It starts with a real person, a real issue, and a real point of view. The AI is then used to help that message travel further, faster, and across more languages than would otherwise be practical.

For Harry, that matters.

His work is physical. It’s difficult. It’s often dangerous. It involves cutting abandoned nets away from reefs and the seabed, hauling waste back to shore, and showing people what’s happening beneath the surface. He has many, many life-or-death accounts of his encounters with underwater ghost nets.

That’s a reality that shouldn’t be replaced by AI.

But it can be extended by AI. Amplified by AI.

A person like Harry can't be in five countries at once. He can't personally record every message in every language. He can't attend every school, every event, every campaign, and every community screening. As much as he’d like to.

But with the right consent, the right process, and the right controls, AI can help him speak to more people without losing the connection to who he actually is.

That's where I think the opportunity lies.

Not in replacing people or creating fake spokespeople.

Not in churning out synthetic content because it is cheaper than doing the work properly.

The real opportunity is in helping real people, real experts, and real campaigners reach audiences they’d never otherwise reach.

Of course, this needs guardrails. Let’s not be naive; there’s a fine line to tread.

Permission matters. Disclosure matters. Human approval matters. Provenance matters. The original material matters. The audience should be able to understand what was filmed, what was edited, what was cloned, what was translated, and what was generated.

That shouldn’t be a restriction on creativity. But it should be the foundation of trust.

This is also why content authenticity standards such as C2PA and Content Credentials are becoming increasingly important. The industry needs better ways to show where media came from and how it was made. And these technologies are becoming more sophisticated by the week - embedded in cameras, browsers, and other hardware and software. In a world of deepfakes and synthetic misinformation, provenance isn’t just a technical bit of meta-data. It’s a credibility anchor that’s part of the bigger story.

For me, the Harry Chan Earth Day project is a useful example of where AI video can go when it’s handled responsibly.

• The ocean waste wasn't fake.

• The ghost nets weren't fake.

• Harry wasn't fake.

• His voice wasn't fake.

• His mission wasn't fake.

AI helped us translate, scale, and distribute the message. But it didn't invent a reality behind it.

That’s a big difference.

The future of AI video shouldn't be about generating more content for the sake of it. We have enough of that already, thanks.

It should be about finding real stories worth telling, capturing them properly, protecting their authenticity, and then using technology to help them reach the people who need to hear them. Stories that in the past wouldn’t have reached such a wide audience.

That, to me, is where AI video becomes genuinely useful: instead of creating a fake reality, we’re helping more people see what is already true.

Tony Jones is a lecturer in Generative AI Filmmaking at the University of Hong Kong and Managing Director of Adaptive Media Partners, a Hong Kong-based AI video and communications company specialising in multilingual video, digital human avatars, and responsible generative AI production. He works with brands, institutions, and regulated organisations across APAC and EMEA to help them use AI to create more effective, scalable, and authentic communications. Visit www.adaptivemedia.ai to find out more.

Glossary

C2PA: An open industry standard that allows creators to provide information about the origin and history of digital media to enhance transparency.

Content Credentials: A tamper-evident metadata standard developed by the C2PA that acts as a 'nutrition label' for digital content.

Avatar or Digital Twin: A high-fidelity virtual representation of a real person that mimics their physical appearance and movements.

Uncanny Valley: The discomfort or revulsion experienced by humans when observing an AI or robot that looks and acts almost, but not exactly, like a real human.

Provenance: The documented history of an asset's ownership or origin, crucial for verifying the authenticity of AI-assisted media.

Generative AI (or Gen-AI) Video: AI models and tools used to create new video content from text, images, or existing video references.

Disclaimer: The content provided in this article is the property of Tony Jones and is shared on Eight PR's website and LinkedIn page for informational purposes only. We do not claim ownership of any content, images, or intellectual property therein. All rights are reserved by the original creator. Eight PR makes no representation as to the accuracy, completeness or suitability of this information. Any reliance you place on this content is at your own risk.