Artificial Intelligence Applications
in TV Industry Workflow

Alan Guedes a.guedes@reading.ac.uk

Who I am?

I am currently Computer Science Lecturer and supervise a CS PhD student.

My research interest lies in Computer Vision & GenAI

I enjoy interdisciplinary, e.g., collaborate with NHM, MERL, UoR Synthetic Media, UoR Digital Humanities COP

Agenda

  • Context of AI in TV Industry
  • AI applications
  • Final Remarks

Context of AI in TV Industry

Streaming Era bring data richness

center


This data richness enabled AI Transformation in TV Industry

Added Values by AI

Added Values of AI Transformation

Operational Efficiency Creative and Market Value
  • Scale & speed: handle large content volumes and enable fast, multi-platform delivery.
  • Cost efficiency: automate repetitive tasks and optimize resources with smarter workflows.
  • Content Quality: improve QC, restoration, adaptation, and aesthetic guidance.
  • Audience value: enhance relevance, accessibility, personalization, and compliance.

It challenging to fully understand current practices and identify research opportunities.

Research Question

How can we map current AI applications in the TV industry and identify emerging research opportunities?

Research Question

How can we map current AI applications in the TV industry and identify emerging research opportunities?

To address it, we:

  • 1. Choose a taxonomy: Owens' TV workflow phases (pre-production, production, post-production, and distribution.)

    2. Reviewed academic literature and industry reports (especially coverage from the International Broadcasting Convention (IBC)).

Owens, J. and Millerson, G., 2012. Video production handbook. Routledge. 2023.

TV Workflow (Owens)

Owens Activities Ideation
Scripting
Planning
Scheduling		 Acquisition
Studio/field ops
Live/remote workflows		 Editing
Adaptation
Localization		 Packaging
Recommendations
Personalization
Advertising

TV Workflow (Owens)

Phase Pre-production Production Post-production Distribution
Owens Activities Ideation
Scripting
Planning
Scheduling		 Acquisition
Studio/field ops
Live/remote workflows		 Editing
Adaptation
Localization		 Packaging
Recommendations
Personalization
Advertising
AI Applications ??? ??? ??? ???

AI Applications in TV Workflow

AI applied to Owens's Workflow

Phase Pre-production Production Post-production Distribution
Owens Activities Ideation
Scripting
Planning
Scheduling		 Acquisition
Studio/field ops
Live/remote workflows		 Editing
Adaptation
Localization		 Packaging
Recommendations
Personalization
Advertising
AI Applications Automated Journalism
Storytelling & creative ideation
Colorization of archives 		??? ??? ???

Pre-Production

Automated Journalism

  • Event detection from newswire and social media
  • Assistant tools for fact retrieval.
  • Newsroom integration for lineup planning, script variants, and compliance hints.

Pre-Production

Storytelling and Creative Ideation

  • In 2016, an IBM composed a 6-minute movie trailer for the horror film "Morgan" by training from 100 trailers

  • 2024, "Dramatron" scaffol Narrative, character and scene generation (LLM + planning).

Phase Pre-production Production Post-production Distribution
Owens Activities Ideation
Scripting
Planning
Scheduling		 Acquisition
Studio/field ops
Live/remote workflows		 Editing
Adaptation
Localization		 Packaging
Recommendations
Personalization
Advertising
AI Applications Automated Journalism
Storytelling & creative ideation
Colorization of archives 		Camera automation (e.g. framing)
Live event analysis (e.g., tracking and overlays)		 ??? ???

Production

Camera automation (e.g. framing)

  • Auto-framing and re-framing guided by face/talent tracking.
  • Shot-type suggestions (wide, medium, close-up) to support operators.

Live event analysis (e.g., tracking and overlays)

  • On-screen graphics from CV and telemetry (speed, heatmaps, zones).
  • Real-time profanity/graphic content flags for compliance chains.
Phase Pre-production Production Post-production Distribution
Owens Activities Ideation
Scripting
Planning
Scheduling		 Acquisition
Studio/field ops
Live/remote workflows		 Editing
Adaptation
Localization		 Packaging
Recommendations
Personalization
Advertising
AI Applications Automated Journalism
Storytelling & creative ideation
Colorization of archives 		Camera automation (e.g. framing)
Live event analysis (e.g., tracking and overlays)		 Super-resolution & upscaling
Colour enhancement & Denoising
Synthetic media generation (e.g., actors)		 ???

Post-Production

Super-resolution & upscaling

  • Temporal SR from SD/HD → UHD/8K with detail preservation.

Colour enhancement & Denoising

  • Auto color-match across shots.
  • Noise/deblock reduction prior to mezzanine encoding.

Synthetic media generation (e.g., actors)

  • Background replacement, inpainting, de-aging and re-timing.
Phase Pre-production Production Post-production Distribution
Owens Activities Ideation
Scripting
Planning
Scheduling		 Acquisition
Studio/field ops
Live/remote workflows		 Editing
Adaptation
Localization		 Packaging
Recommendations
Personalization
Advertising
AI Applications Automated Journalism
Storytelling & creative ideation
Colorization of archives 		Camera automation (e.g. framing)
Live event analysis (e.g., tracking and overlays)		 Super-resolution & upscaling
Colour enhancement & Denoising
Synthetic media generation (e.g., actors)		 Recommendation
AI assistants
Targeted Advertisements
Multilingual & Adaptation

Distribution

Recommendation

  • Hybrid recommenders (content + behavior) to mitigate cold-start.
  • Moment-level retrieval (faces/objects/topics) to power semantic search.
  • Diversification
  • Measurement fatigue control.

AI assistants

  • Natural-language queries over catalogs (“the episode where…”).
  • Conversational browsing; hands-free accessibility features.

Targeted Advertisements

  • Contextual + audience-based decisioning with brand-safety filters.
  • Creative optimization (A/B/n) and attention/quality signals.

Distribution

Multilingual & Adaptation (e.g., dubbing)

  • ASR-driven captioning and automatic timing.
  • Machine dubbing with licensed voices; lip-sync and prosody control.
  • Style-preserving translation
  • Example: TrueSync on the movie "Fall"

Final Remarks

Final Remarks

Owens' TV Workflow can be a usefull perspective to understand AI in TV industry

AI adoption in TV Industry is uneven

  • Mature in distribution (recommendations, ads) and post-production (enhancement, localization),
  • Exploratory in pre-production and production.
    • potential for integrated AI within human editorial and technical decision-making.

Opportunities for Researchers

TV industry high profit and competitive market, so very cloded

Academic can promote

  • open, reproducible evaluation frameworks and shared testbeds .

  • interdisciplinary collaboration

  • Benchmarks and datasets specific to broadcast and media workflows.

    • Interoperability & metadata standards
    • Quality metrics for aesthetics, accessibility, and editorial integrity.

  • bias-free human–AI collaboration in creative/editorial decision-making and its evaluation.

Thank you

Alan Guedes, a.guedes@reading.ac.uk