At our International Quorum of Motion Picture Producers annual conference in Atlanta last October, one of the panels looked at the legal complexities of GenAI. During the Q&A, one of the excellent, experienced filmmakers captured our industry's frustration perfectly:

“When will these companies pay us for the intellectual property they've taken and which now threatens our livelihoods?”

“Well, there’s so much money involved, they'll figure something out. There’s a deal to be done.”

“How exactly will they figure it out? When will we see compensation?”

[Silence.]

“They’ll figure it out.”

This exchange encapsulates why, despite the fevered excitement around AI video and text, our industry remains a little ambivalent. It also helps to explain why so many enterprise brands are hesitant to fully embrace GenAI technology.

The Emperor's New Clothes

The theft of intellectual property by the companies that have trained our future AI overlords is arguably one of the great unspoken crimes of our age. Whether it is legally theft - or not - is currently unsettled law. Living in San Francisco, I am familiar with the hype cycle of technologies, where no one wants to question the technical emperor’s new clothes for fear of being seen as behind the times. Almost the worst thing you can be here. Almost.

Human vs. Computer 'Reading'

From my point of view, there is a fundamental difference between the imperfections of a human reading/watching/remembering information and then replicating it, and the perfect replication of a computer system doing the same. These models are not 'reading' in the way a human does.

Added to the fact that if you were to ‘read’ ChatGPT’s source code and then copy it, they would sue you within an inch of your existence. Sam Altman complaining that the Chinese model, DeepSeek, had copied them was so rich I could hardly believe it. Although, of course, shame is now dead.

Slowwly Slowly Catchy Monkey

All this explains why, despite all the cacophonous noise around GenAI text and video, so few of our clients are genuinely keen to use it. In fact, more of our clients have specifically told us not to use it, rather than the other way around. Now, of course, enterprise brands are cautious out of necessity. With billions at stake and reputations earned over decades, Fortune 500 companies view groundbreaking tools like GenAI video with caution. That hesitation makes sense.

So what are some of the reasons that drive this prudence, when, to judge by the breathless coverage of AI, the returns seem so great. Here are six of them…

1.    Regulatory Minefields

Regulatory uncertainty poses significant challenges. Europe's upcoming AI Act, effective from August 2025, mandates explicit disclosure of synthetic or AI-generated videos. Global brands operating across multiple territories face potential compliance complexities, making heavy investment in GenAI video risky until clearer guidelines emerge.

2.    Legal Battles Yet to be Settled

Unresolved legal disputes create additional uncertainty. Major AI companies - Google, OpenAI, and Stability AI - currently face high-profile lawsuits over unauthorised use of copyrighted material in training their AI models. Until courts clarify whether these practices constitute fair use, corporate legal teams remain understandably wary of inadvertently becoming caught up in litigation.

3.    Safeguarding Brand Reputation

Brand safety concerns remain paramount. Enterprise businesses have witnessed negative backlash against AI-generated campaigns from major brands like Coca-Cola, which are criticised for appearing impersonal or culturally insensitive. One dodgy GenAI campaign could result in lasting reputational damage, outweighing any potential cost efficiencies.

4.    Protecting Sensitive Data

Data privacy and confidentiality concerns are acute for enterprises. Many corporate videos include proprietary or sensitive information. Uploading confidential content into GenAI systems risks inadvertent data leaks or unauthorized exposure. Given stringent security requirements, enterprise-level brands understandably hesitate to adopt workflows that might compromise their own intellectual property.

5.    Mixed Creative Quality

Despite the incredible advances we’ve seen, there are quality issues with AI outputs. Visual continuity, crossing the uncanny valley and adherence to brand standards are still issues. It takes a lot of work to get videos to look right – this offsets the perceived advantages of speed or cost efficiency.

6.    Ethical and Social Accountability

Despite the recent retreat from DEI, ethical and social considerations continue to significantly influence enterprise decisions. Commitments to ESG goals, union relations, and responsible AI practices require enterprises to carefully evaluate potential implications. The perception of AI displacing jobs or producing culturally insensitive content could trigger significant backlash, particularly in highly regulated or unionised environments.

In Summary

Of course, our clients will overcome these concerns – some already are. We’re all feeling our way forward, experimenting with clearer disclosures, hybrid human-AI workflows, and promising pilot projects. As I have written before, these tools will empower businesses that show the most curiosity, creativity, and agency. That is a genuinely exciting prospect. But it has to be done carefully and sensibly, and the people who have had their life’s work fed into training these models need to be made whole. But that’s another blog.

Have a good week.