Seeing the wood for the trees: Globecast explores the potential of AI on sports broadcasting
By Steve MacMurray, Globecast digital media development manager
Without a doubt, the topic that’s been highest on the agenda at the most recent broadcast conference events is artificial intelligence (AI). It has risen from a subject for discussion by curious tech teams to suddenly become a focus for company leaders.
Equally as noticeable is the huge gap between what’s being discussed and reasoned, and actions being implemented right now. What are the next steps for the broadcast industry? There’s a widespread concern – even anxiety – about the impact of AI on our working lives. Businesses want to achieve a level of preparedness for it, with responsibility and safeguarding at the forefront of their thinking. Yet, simultaneously, there’s an appetite for incorporating the enormous benefits AI promises as part of anyone’s technology transformation strategy.
The advancements in GPU-based computing, language processing, big data and cloud computing economics have made AI more accessible, affordable and immediate than ever before: it has been on everyone’s lips, from the medical community’s application in finding new antibiotics, to politicians warning of the potential risks to humanity. Closer to our industry, the recent Hollywood writers’ strike underlined the palpable threat that generative AI casts over the creative arts. Clearly, there are hugely positive gains to be had, fearful losses to be made, and a huge grey area in between to safely navigate.
AI in broadcast
Looking at it from the perspective of a service provider, what exactly are the practical uses for AI in the broadcast market? Again, the spectrum’s very broad, and we’re only at the beginning of the learning journey. Generative AI, viewer engagement, metadata management, optimisation of workflows, immersive experiences: the applications are legion. With every organisation going through a digital transformation, how do we blend the best of our traditional broadcast tech stack with these powerful changes?
Generative artificial intelligence (also known as generative AI or Gen AI) is artificial intelligence capable of generating text, images, or other media, using generative models. It’s a much-discussed concern when it comes to script writing, deep-fake actors, overall content production – but that’s not directly relevant to the subsector in which we operate. For us, it’s more about the efficient automation of tasks and identifying where additional efficiencies may lie. The aim is to optimise complex processes and automate them via trusted AI, freeing up human expertise to concentrate on other possibilities. The idea is a natural extension of what we’ve always done in our business: human experts working alongside hard-working technology.
AI is already being used for logging and metadata insertion in sports content. There are a number of organisations offering this service, ingesting live sports feeds and using AI to log every noteworthy incident. AI is given time to ‘learn’ a sport and the way it works, resulting in accurate and speedy content enrichment. Looking at the metadata side, it’s incredibly fast and cost effective in terms of incident logging at a complex level – every corner, goal chance, foul incident, VAR decision and so on.
The result is, whoever is creating the story, AI has indexed all that content and made it meaningfully searchable. It can work as a crucial tool in helping pull the story together, with agility and speed. Again, this is using AI to maximise human creativity in doing the heavy lifting, delegating labour-intensive tasks, to help create a greater whole.
In applying this process to years of archived material that all media organisations have, not just sports content, the use of AI to add value is a compelling one. Content can be unlocked, opening it up for monetisation, or remonetisation, empowering content creators to develop new stories or innovative programme ideas with AI doing the leg work of panning for content gold.
Looking ahead
The applications go much further, playing into areas such as production efficiencies, camera automation, content gamification, in-play betting, live match data analysis and fan engagement. We’re only just beginning to see the practical applications from both from a broadcaster perspective and a club or event owner standpoint.
What’s also interesting to look at is how AI can positively impact a broadcast operational environment such as ours. As an industry, I think we’re only at the very start of a rapid period of development and the discovery of the possibilities this holds. For multiple, complex live feeds created in the cloud, which require replication on a weekly basis for a particular sports broadcaster or rightsholder, the question is, how much of that work can be performed by AI, once it’s fed the initial understanding of what’s required? The aim is not to design out the human expertise – either at the initial stage or the live monitoring point – but to trust the repetitive, complex task elements to be handled by AI. Of course, it can then also help with troubleshooting issues or even identifying optimised workflows. AI will have a dramatically significant role to play in our industry – many roles, in fact – as content volume and consumer demand grows, in the face of cost efficiency and sustainability goals.
Even going beyond the impact on operational workload in the broadcast MCR, there is exciting progress taking place in the world of media processing – with delivery applications ranging from automated encoder configuration/codec optimisations, on to more advanced developments such as super-resolution – leveraging AI to increase image quality.
Going to the far end of the transmission chain, the fragmentation of subscription services and the navigation issues troubling audiences is an issue that AI can also help solve. Audiences find it awkward to switch around different platforms, and they clearly prefer the ease of a single-entry point. This introduces the notion of a ‘digital concierge’, which is potentially very appealing. At a simple level, if you’re not a pay TV subscriber but you are interested in a particular sporting event or match, this service could let you know when and what the possibilities are for viewing. Accurate recommendations, and easier discoverability, are powerful tools to drive the viewer to the live moment.
But let’s not give way to panic about fearing human obsolesce! Our creative input remains vital. Emotion is a fundamental of engaging storytelling, and passion and humour are essential to a successful narrative. I’ve seen a recent demonstration of an automated AI football match commentary, and it immediately felt dull and soulless. Admittedly, it was very clever in the way it was instantly reacting to the match action and crowd reaction, but it was a million miles away from the unique creative and passion engagement delivered by a professional commentator.