Where can AI be used safely?

Artificial intelligence is reshaping the world of broadcast. New tools emerge weekly, promising to automate workflow and accelerate production. It’s exciting, but complex. We are living in a time where technology sprints, regulation walks, and liability crawls. 

The growing gap between innovation and regulation is why the industry must be cautious. Today, the safest, most responsible, and most commercially meaningful use of AI is enhancement - not replacement. 

Copyright: the core constraint 

Copyright around AI-generated content is one of the most debated legal issues globally. The truth is straightforward: in most jurisdictions, a work must involve genuine human contribution to be protected. AI, by definition, struggles to meet that standard. 

In the EU, the criteria are explicit. For something to qualify as a copyrighted work, it must pass a four-step test: Production, Human intellectual effort, Originality, and Expression. AI falls short at step two. Without human intellectual effort, it cannot satisfy the legal definition of a “work.” 

The UK is slightly more open to the conversation, but even under the Copyright, Designs and Patents Act 1988, human involvement remains essential. It is a highly contested space. 

Will AI output ever be fully protected by copyright? Possibly - but certainly not in the immediate future. We are still debating whether AI can replace human jobs. We are far from a world where AI-generated scenes or dialogue are recognised as original creative works under the law. 

Authorship is a spectrum - not a checkbox 

Courts and regulators analyse AI-assisted works not based not on what is produced but how it is produced. They examine: 

 - What humans contributed (story arcs, emotional logic, scene structure, character identity) 
 - How much control humans exercised (iterative refinement, direction, selection) 
 - How the AI system was positioned (as a rendering tool or a creative substitute) 

Two outputs that look identical may have radically different legal outcomes depending on the process. 

For example, a two-second video of two people arguing can be unprotectable if the user simply writes: “Create two people arguing in the street.” In this case, AI invents most of the creative elements. 

But if a creator specifies the narrative, emotional tone, visual language, and structural elements of the scene with a highly detailed prompt (e.g., specifying clothing, lighting, camera position, and emotional tone), the AI system functions like a technical tool executing a pre-conceived creative vision. This degree of human contribution could amount to authorship. 

This does not mean all AI-assisted works are, or will be, copyrightable. A generic prompt may not qualify; an intentional, artistic set of instructions might. Case-by-case assessment is inevitable. 

The way forward - AI for enhancement 

Amid legal uncertainty, enhancement is where AI delivers real, safe, and immediate value. Enhancement respects existing rights because it doesn’t create new materials; it simply improves what already exists. 

Upscaling, colour correction, noise reduction, and quality harmonisation - these enhancements protect both the artistic intent and the copyright of the original work. 

The commercial impact is undeniable. Platforms like YouTube reward quality. A better video means stronger ad performance and more engaged audiences. Older content becomes relevant where audiences expect high-fidelity experiences. By enhancing the viewing experience, you’re enhancing the value. 

AI to navigate audience behaviour 

Another area where AI excels is audience analysis. No human team can read thousands of comments across multiple platforms - but AI can, and it can summarise what people resonate with. 

For live series, this insight can shape production in near real-time. Production teams see fan-favourite characters, moments that trigger emotion, and where attention drops. They refine storylines with real-time intelligence. 

AI is also starting to support monetisation. Tools identifying optimal ad-break placements help retain viewers and maximise revenue. This is not replacing creativity - it is strengthening the business engine behind it. 

The regulatory challenge - liability 

Perhaps the most serious unresolved issue is liability. When AI fabricates or distorts information, who is responsible? The model creator? The platform? The user? 

Courts are not yet equipped to answer that. Regulation must evolve, supported by academic guidance and judicial understanding of AI systems. This will take years - all while AI evolves every day. 

The strategic position for the industry today 

AI is no longer optional. It is a structural force reshaping production, monetisation, and audience behaviour. 

But success belongs to those building hybrid workflows: 

 - Humans invent, define, and direct.
 - AI amplifies, accelerates, and enhances. 

Creativity remains human. And in today’s legal reality, that is not only a cultural truth - it is a commercial advantage. 

The leaders of this era will not ask whether AI can replace humans. They will ask how far human ambition can scale when AI becomes its multiplier. 

Ceyda Sıla Çetinkaya is COO of Merzigo