AI technology in broadcasting - is it still all hype or the next big thing?
As a topic, Artificial Intelligence (AI) has certainly seen plenty of hype. Gartner’s ‘hype cycle’ is a well-recognised measure of new technology adoption curves and, when it comes to the subject of AI in broadcasting, Gartner has suggested that:
“AI is almost a definition of hype. Yet, it is still early: New ideas will surface and some current ideas will not live up to expectations.”
That was in 2018. Since then, AI awareness and adoption has advanced beyond levels 1 (‘Early AI interest with a risk of overhyping’) and 2 (‘AI experimentation, mostly in a data science context’) on Gartner’s AI Maturity model. In fact, there are many organisations that are already at level three, defined as ‘AI in production, creating value by e.g. process optimisation or product/service innovations’ and there are practical applications emerging today.
Investment in AI is on the up
This should come as little surprise, as investment has been increasing. The IABM’s bi-annual Buying Trends survey from late 2018 showed a huge boost in AI and Machine Learning (ML) adoption in the broadcast and media industry, up from 2% to 13% in a mere six months (April to September 2018), with 68% of organisations stating they are likely or very likely to deploy AI in the next 2-3 years.
But what are the real world use cases? AI’s affinity with automating repetitive tasks is proving a boon in the broadcast sector, resulting in a wealth of applications that can refine complicated workflows, deliver usable and accurate speech-to-text, automatically generate metadata and create actionable data from analytics.
An IBM-powered player spotlight
One real-world example was on display via Fox Sports use of IBM’s Watson in the FIFA Women’s World Cup. Fox used the AI tech to build a segment called "Player Spotlight", which used Watson to create statistical analysis by ingesting data from tournament data provider Opta.
Fox Sports commentators could query the tool directly, and use the highlights and analytics on-air. Fox has used a similar architecture in the past, during the 2018 FIFA Men’s World Cup, where Watson was used to analyse archive footage for goals, red cards and individual players, which users could in turn search for and view.
Sports highlights is an area that AI is rapidly becoming a defacto tool. At Wimbledon 2018, an IBM Watson powered system was used to clip more than 75,000 individual points from some 1,000 hours of footage, which had the happy result of shortening the turnaround time of the highlights package from 45 minutes to 5 in the process.
Moving into the cloud
One of the major migrations in broadcast recently has been the move to the cloud, a move that not only reduces the requirements on individual companies for on-premise hardware investment, but also enables the use of in-cloud software, delivery infrastructure and of course almost infinite scalability and resiliency into the bargain.
While cost is, of course, a factor, this move has been and will continue to be a strong driver for initial adoption of AI and ML services. Amazon Web Services, for example, has a ‘Free Tier’ that offers free ML services including 5 million text to speech characters per month, speech to text at 60 minutes per month and image recognition. The growth of AI as a Service (AIaaS) is highly likely to accelerate, especially for mid-tier and smaller broadcast players.
The consumer’s voice
Of course, it is not just B2B applications that are proving fruitful avenues for broadcasters deploying AI, as clearly exemplified by consumer-facing brands such as Amazon and Google.
Not only is the power of semantic search increasingly important to entertainment consumption, but voice search is rapidly resolving the conundrum of the convoluted EPG and restricted devices such as numeric keypads on the average home remote control. Voice-based search is rapidly becoming one of the most important digital navigation methods, and of course, is highly device and platform responsive.
For example, the first live broadcast of French Open matches in 8K might have hit the headlines for the use of 5G, or the groundbreaking quality of the content. But there was also an accompanying smartphone virtual assistant helping visitors navigate Rolan Garros.
Mobile operator Orange (which partnered with the French Tennis Federation and France Télévisions for the event) created a text and speech-recognising AI that answered French Tennis Open questions posed by visitors in real-time.
The future of AI in broadcasting
Gartner’s Level 4, described as ‘pervasiveness’, is where ‘AI is pervasively used for digital process and chain transformation, and disruptive new digital business models.’ It’s arguably just around the corner, as late adopters take advantage of low-risk AIaaS services, and take the lead from early adopters and tech pioneers.
Additionally, as ROI becomes clearer, more targeted investments will follow from latecomers and first movers alike. Projects such as Channel 4’s Contextual Moments, which aims to deliver contextual advertising for linear TV at scale, will continue to proliferate.
What is absolutely certain is that IBC2020 will track the progress of AI in broadcast, as well as examine the biggest catalysts and inhibitors to the market. As technologies advance towards Level 5 on Gartner’s AI Maturity Model (‘Transformation’), we are moving towards a future where ‘AI is a part of business DNA.’
Explore this and other groundbreaking technologies as the world's most influential media, entertainment and technology show returns 11-15 September 2020. Don't miss out on any news and updates, register your interest for IBC2020 today.