Accedo – Can AI support the transition towards a more sustainable video ecosystem?

Accedo – Can AI support the transition towards a more sustainable video ecosystem?

IABM Journal

Representing Broadcast & Media Technology Suppliers Worldwide
Articles covering a range of key topics and themes
In depth analysis of the latest trends

Accedo – Can AI support the transition towards a more sustainable video ecosystem?

Tue 16, 04 2024

Can AI support the transition towards a more sustainable video ecosystem?

François Polarczyk, Sustainability Director, Accedo

 The OTT industry has undergone some major changes over the past few years. Market growth slowed somewhat compared to previous years and video providers have broadened their monetization strategies and shifted focus from subscriber growth to profitability. Despite this, the OTT video industry remains buoyant; according to analysis by Statista, the industry is projected to show an annual growth rate of 6.30% between 2024 and 2029, to reach US$429.40bn by 2029. This change of focus towards profitability is driving service providers to provide a better experience for viewers and optimize their services. However, there is a need to balance this drive for profitability with the industry-wide need to transition towards a sustainable video ecosystem.

This is a complex challenge, and we know that it will take considerable innovation. Might the rapidly accelerating innovation we’re seeing around AI help the industry to pave a sustainable path? Already we’re beginning to visualize the impact that AI might have on content creation, OTT design, engineering, delivery and consumption. How might AI support the industry to respond to the challenges of today to make the transition towards a more sustainable video ecosystem tomorrow? Clearly, the potential for AI to provide positive change is huge.

 Reducing impact of content creation

Content production is responsible for producing huge amounts of carbon dioxide. Environmental organization albert estimates that the production of a TV show typically produces tens of tons of CO2 while a feature film might produce tens of thousands of tons. A huge amount of resources go into set production, not to mention the carbon emissions from energy consumption. If more sets can be created virtually with the help of generative AI, fewer physical sets need to be created which would reduce resource consumption and waste generation. There is also potential that virtual sets could lessen the need for location filming, which would reduce emissions from transporting people and equipment, as well as reducing emissions produced by generators.

AI could also be used to enhance sustainability in OTT content creation by helping producers determine which production processes are the most efficient. This can lead to more efficient use of energy, making production methods more sustainable. Content producers are already using AI to streamline post-production and localization by automating time-consuming tasks. As more AI powered tools come to market, we’ll no doubt see higher levels of automation across the content supply chain, which ultimately will contribute to making the industry more energy efficient.

Optimizing software development and design

Just as AI-driven solutions are already playing an integral part in the post-production process enabling content to be produced quicker, generative AI is also being used by software developers to optimize the OTT video service development process. Already we know that AI can help developers to write code more efficiently by automating repetitive coding tasks and speeding up the review process. Might AI also help developers to optimize code so that the computational requirements of the video software are lowered? It’s reasonable to hope that AI can help not only with making the practice of software development itself more sustainable, but also with helping to reduce energy consumption further along the value chain.

There’s also potential that AI may also positively impact the product design process, to enable services to be designed in a way to ensure that both user experience and energy efficiency are always optimized and balanced in harmony with one another. Perhaps AI-powered tools could help individuals and households understand how their viewing behavior and decisions affect their carbon footprint.

Delivering content in the most energy efficient way

Content needs to be delivered to viewers in the most energy efficient way possible. We’ll likely see AI-powered video compression algorithms helping to reduce the size of video files, without compromising quality. This has the potential to help lower bandwidth requirements and reduce energy consumption during data transmission.

Additionally, by using AI algorithms to analyze large datasets around usage and network conditions, video providers will be able to better understand energy consumption patterns and identify inefficiencies in real-time in order to dynamically optimize content delivery.

AI-powered tools are already starting to come to the market that allow Content Delivery Networks and even data centers to optimize network efficiency while at the same time reducing energy consumption. AI can greatly help with ensuring resource intensive workloads are predicted by AI and therefore help with scalability and elasticity. Video providers will be able to automatically optimize the routing of data through the network, choosing paths that consume less energy. This will be particularly valuable for large Content Delivery Networks where efficient routing can lead to substantial energy savings.

Content discovery is another issue that impacts on the energy consumption of streaming. As the volume of available content increases, people spend more and more time looking for something to watch, which means TVs are powered on for longer periods of time. One concept that we’re working on to address this issue is to incorporate AI-powered content discovery that would reduce energy by dimming the entire screen while also getting the right content in front of the viewer, faster.

Looking ahead

While there is little doubt that AI can help tremendously with efficiency, training and running large AI models does itself consume huge amounts of energy. A recent study concluded that by 2027, global electricity consumption relating to AI could increase by 85 to 134 TWh annually, which is comparable to the annual energy consumption of a small country. Although significant, if this energy can be generated from renewable sources, then that obviously goes a long way towards reducing its impact.

There are indicators that AI may well be a technology that could help on both those fronts: firstly by helping video providers enhance UX and service quality to deliver more value to consumers, and secondly by enabling optimization at all stages of the value chain to improve efficiency and reduce energy use.


Search For More Content