We’re amid a streaming and broadcasting revolution, with an explosion in the number and availability of over-the-top (OTT) platforms like Netflix, Hulu, Disney+ and Amazon Video. Accelerated by consumer needs and preferences during lockdown, innovation is well underway, as each platform looks to gain a competitive edge. While we’re seeing innovation that utilises a range of technologies and functionalities, here we focus on artificial intelligence (AI) and machine learning (ML) to explore how OTT content production and delivery is being transformed, as well as how innovators can best patent technologies in this space.
AI and patent law
First, here’s a brief overview of AI in terms of European patent law.
How are AI inventions treated by the European Patent Office?
AL and ML inventions are viewed as (computer implemented) mathematical methods by the European Patent Office (EPO) and are therefore governed by the same provisions. While computer software and mathematical methods are commonly believed to be excluded from patentability, this isn’t always the case.
Article 52(2)(c) of the European Patent Convention (EPC) excludes computer programs and mathematical methods “as such” from patent protection. However, it can be possible to protect AI and ML inventions if, above the usual requirements around novelty, inventive step and industrial application, you’re able to identify the technical purpose of the AI invention and draft the patent application around it. Alternatively, you may be able to bring out a specific technical implementation of the AI within the invention to demonstrate that it works in a different way to pre-existing technology. There are some specific EPO guidelines around this.
Can you patent a new use involving existing AI concepts?
Importantly for OTT innovators, you can gain patent protection in Europe for a new use or application involving existing AI — provided that the AI serves a specific technical purpose. With AI- and ML-related technologies set to boost content production and revolutionise the way we access video content, this point is crucial.
The EPO guidelines helpfully provide a (non-exclusive) list of examples where mathematical methods could be considered as addressing a technical purpose. Some of those examples that are relevant to the use of AI and ML for OTT are:
- the classification of digital images, videos, audio or speech signals based on low-level features (e.g. edges or pixel attributes for images)
- digital audio, image or video enhancement or analysis (e.g. de-noising, detecting persons in a digital image, estimating the quality of a transmitted digital audio signal)
- separation of sources in speech signals; speech recognition (e.g. mapping a speech input to a text output)
On a related note, it’s important to bear in mind that there are already a number of patents in the OTT and AI spaces. Therefore, prior to launching a new OTT product or service that uses AI, innovators should consider the benefit of keeping an eye on third party patents that are (or will soon be) in force in their launch countries, as any overlaps could cause problems.
How will AI revolutionise OTT platforms?
We think that there are at least three main areas in which AI innovation is set to transform the OTT experience.
1. Content production
The first of these is to make content production faster and much more efficient, reducing costs and time-to-market — important in a world that is currently dominated by video consumption. A great early example of this took place in 2016, when 20th Century Fox teamed up with IBM to develop the world’s first “cognitive movie trailer” for its horror flick ‘Morgan’. This involved training an AI using spectacular moments from other horror films and trailers to put the new trailer together in a way that most appeals to horror fans.
From IBM: “There are patterns and types of emotions in horror movies that resonate differently with each viewer, and the intricacies and interrelation of these are what an AI system would have to identify and understand in order to create a compelling movie trailer. Our team was faced with the challenge of not only teaching a system to understand, “what is scary”, but then to create a trailer that would be considered “frightening and suspenseful” by a majority of viewers.”
The AI (with editing help from humans towards the end of the process) achieved this feat in just 24 hours — without AI assistance, producing a movie trailer usually takes between ten and 30 days.
Five years on, the ability of AI and ML to sense, learn and adapt has improved even further. Content production can be automated in many ways, reducing the time and manpower required. AI can learn from error patterns and speed up data processing, and (as with the early example of the Morgan trailer) create new videos in real time from content libraries, perfectly optimised for release on each different OTT or social media platform, without the need for any human assistance.
If such giant leaps continue to be made in content production, perhaps someday we could even see individual content creators producing Hollywood-quality movies with the assistance of AI.
For consumers, this is all great news — since it will be easier and quicker for OTT platforms to produce new video content, they will be able to create more and higher-quality content. With Netflix’s revenue reaching $25bn in 2020, the future earnings potential for OTT platforms seems endless.
2. Content delivery
The second area of transformation is in how content is delivered to end users. We’re already seeing plenty of movement here across the board, as each OTT platform looks to gain a competitive advantage in what has become an incredibly saturated market.
Again, we can look to Netflix as a leader in this space. Its recent patents include those directed to AI- and ML-based recommendation systems, based on user preferences, which ensure fast and precise recommendations with a high degree of personalisation. It also holds patents for technologies that optimise how content is presented to the viewer, enhancing stream quality while reducing buffering and data loss.
Such innovation shows that OTT platforms are now competing in terms of their technological capabilities as much as they are with content. If one platform provides a far smoother viewing experience than the other no matter which device you’re viewing on, with better recommendations that make it easy to find new content, they may attract far many more subscriptions.
The key question here is whether inventions in the content delivery space can be patented. While Netflix seems to have a strong approach here, patent offices have often looked at searching, tagging and organising software methods with suspicion unless the invention is limited to a specific purpose or you can pinpoint exactly which function of the AI is being used to improve a certain aspect of content delivery. It’s important not to fall into the trap of trying to patent a process that simply automates something a human could do mentally to achieve the same result — instead, you must zero-in on the function or feature that solves a technical problem (and that results in an improved process). The trick here is to not claim the invention too broadly (however tempting it may be) and clearly explain not just what, but also how, the improvement claimed is being implemented and/or used.
3. Blending OTT platforms
The third clear way that AI is changing the OTT game is through the blending of platforms. From Instagram to TikTok and Amazon to Disney, all types of social media and streaming platforms are patenting AI-based inventions. This implies that we could be seeing the start of such platforms trying to combine applications and functionalities together. YouTube is a prime example — it hasn’t been simply a video viewing platform for a long time — instead, it’s a video editor, social network, message board and much more.
There are also new entrants to the OTT space, enabled thanks to advances in AI-based technologies. MediaKind Engage is one such challenger — a cloud-based platform set to deliver “extraordinary Direct-to-Consumer fan experiences at scale”, offering the integrated ability to produce video, stream video and engage audiences. It aims to target a wide audience of sports clubs, broadcasters and content owners, providing the opportunity to leverage OTT technologies to expand the reach, scale and reliability of video content streaming. Ultimately, a platform like this acts as a one-stop shop for content production, publishing, distribution and audience engagement. Yet it’s not the only platform that could likely achieve success with an integrated model — imagine how the user base of social media platforms, combined with the content production and streaming capabilities of OTT platforms, could change the game completely.
UX drives OTT innovation
Wherever the market heads, what’s clear is that AI is revolutionising OTT technology. Innovation among platforms is rife as diversification efforts increasingly focus on the user experience, rather than simply the types of shows and movies available to watch.
To protect computer software and mathematical methods you need real expertise — not just in terms of truly understanding the subject matter, but also in terms of experience in dealing with patent offices around the world.
Expertise in AI, ML and software patents
We have a rare combination of experience and depth of understanding, having drafted and prosecuted a significant number of software patents (with a focus on those relating to cutting-edge technologies such as blockchain, cybersecurity and extended reality) and possessing an expert team that consists of both mathematicians and pure computer scientists, enabling us to truly comprehend the underlying technology behind AI-related patents.
The success of a patent application in such a complex field depends on the quality of the draft — get in touch with us to discuss your invention.
