Many new company presentations in London nowadays include a section when a member of the audience, generally a senior creative or occasionally the client, asks explicitly if any of the presentation’s video was created using generative AI. That question wasn’t common a year ago. These days, the response has the power to make or break a sale. Before anyone even considers asking, a few of the city’s larger independent agencies have begun to clearly state in their pitch credentials that they do not use generative videos. It now serves as a positioning statement. created by humans. Everything.
Clients’ resistance has been more significant than many in the business anticipated. AI limits are now directly written into production contracts and creative briefs by major international brands, including consumer products corporations, financial services firms, and automakers with strong legal teams. The reasons differ from client to client, but they generally revolve around two issues: a reputational risk from being linked to content that consumers, or worse, journalists, may someday recognize as machine-generated, and legal exposure from training data that no one can properly audit.
For these businesses, the issue of copyright scraping is real. It’s possible that their proprietary creative assets, image libraries, and historical campaign material wound up in training datasets without their permission. This leads to a liability discussion that brand legal departments are unwilling to take lightly.
The ethical issue is exacerbated by the quality issue. Although generative video technologies have rapidly advanced, they remain unreliable in many areas that are crucial for commercial production. Human faces continue to be the most common source of failure: the tiny inaccuracy of generated skin, the way eyes don’t quite track correctly, and the subtle temporal anomalies that untrained eyes perceive but are unable to identify. The text in the created video is often inaccurate or warped. Hands, fabric movement, and liquid behavior are examples of tiny physical features that still generate outcomes that appear good in a still frame but crumble in motion. The tools aren’t really production-ready for the majority of what they’re supposed to deliver, which is problematic for agencies whose creative reputation depends on precisely that level of physical correctness.
Although it runs beneath both the legal and quality concerns, the talent aspect of this uprising typically receives less attention. Pipelines—junior editors, entry-level motion designers, and assistant directors who spend years honing their skill before being trusted with more—are the foundation of London’s creative agencies. The institutional expertise that keeps senior creative departments competitive is likewise produced by those pipelines. Agencies that eliminated those positions in favor of generative tools in 2023 and 2024 are starting to realize that the abilities they lost are difficult to replace and that the tools did not completely replace what those individuals were capable of. The industry has been monitoring a staff exodus that isn’t solely due to layoffs; some experienced talent is departing because the workplace no longer feels like a place where skill is appreciated.
There is additional pressure coming from a different angle because to the larger legislative framework in the UK. Compared to other European governments, Parliament has been examining AI copyright laws with an unprecedented level of seriousness, and the creative industries have been successful in presenting their case in that context. In committee rooms, the claim that generative models are what some practitioners refer to as “glorified plagiarism” on a commercial scale has gained traction. It remains to be seen if this results in legislative action that genuinely alters the economics of generative video, but the political risk for companies utilizing these techniques has significantly increased just by virtue of the discussion taking place at that level.
There’s a real sense that what’s forming isn’t just simple technophobia as the London creative industry works thru this. AI tools are widely used in research, scripting, and pre-production by many of the same agencies who reject generative video. The line they’re drawing is more precise; it’s about the final frame, the output that faces the audience, and the object that a client compares their brand to. As the technologies get better, it might be more difficult to retain that distinction, but for now, the industry has chosen to set its flag there. For the time being, the clients appear to concur.
