The market in pirate, ad-funded sport, TV and film content has grown in tandem with the legitimate streaming and pay-TV business and has predictably surged during lockdown. But custom AI innovation has given a form and a face to this previously nebulous market, as well as arming rights owners and advertisers with the tools to take action.
Recommended Blog: Solving for the Browser Imbalance in 2021
Like any efficient black market, the illegal streaming business is both everywhere and nowhere: easily available to those seeking to consume its borrowed wares – whether that’s live sporting events, TV, movies or music – but shadowy and nebulous to those who attempt to pin it down.
It’s easy to imagine, in fact, that online content piracy, with its viral marketing methods based around social media and search, is a small-time concern. But in practice, fueled by its ability to attract legitimate programmatic advertising to its domains, we calculate that internet content pirates generate more than £1 billion in ad revenue globally.
Nor is that earning power purely collective: each of the UK’s top ten most-visited pirate websites makes close to £20 million a year from advertising, much of it from blue-chip brands. Organised criminal operations are making serious money off the back of someone else’s content and the advertising it attracts, and well over a year of lockdowns has heralded rocketing consumption of IP-infringing material, making this parasitic business more lucrative than ever before.
Anyone with a web connection can catch a fleeting glimpse of content piracy in action, but mapping its entirety – its shape, its size and its practices – is now well beyond the ability of unaided humans. Surveying the universe of piracy – analyzing some 3 million at-risk sites every day, as well as 50 app stores and more than 400,000 daily domain registrations and app submissions – requires AI-driven data analysis on a massive scale.
Essentially, the tracking process involves detecting commonalities between known pirate publishers and previously unknown ones. That allows us to identify correlations and trends, as well as verifying IP-infringing content, tracking the advertising being served to those locations and measuring their popularity.
This technology has been developed over eight years and distils extensive experience in pirate site analysis into a continually updated fusion of state-of-the-art machine learning, predictive algorithms and computer vision – including image recognition – to score media for piracy risk across multiple digital ecosystems.
We use infringement analysis combined with contextual, sentiment, infrastructure and ad-bidding assessment to enable granular scoring of hundreds of attributes. White Bullet has processed some 6 trillion ads and growing, and responds in 2 milliseconds to real-time ad bid requests.
Rigor and scale are vital in all of this, because the illegal content ecosystem is enormously dynamic – its most popular domains constantly ducking out of sight and re-emerging in new guises and locations in order to stay ahead of detection.
The pirate ecosystem is also in the midst of a major shift from web domains into apps, and monitoring efforts have had to move in the same direction. The custom AI solution this demands is necessarily subject to constant refinement, but is now capable of tracking and analyzing the app environment, which until recently was largely opaque and beyond the range of such efforts.
Since the advent of P2P music sites such as Napster at the turn of the millennium, the existence of pirate content online has never been any secret, but piracy’s latterday scale and the shape-shifting abilities of its exponents – not to mention their commercial nous – are something to behold.
Certainly, the brands whose advertising budgets keep this pirate market afloat are routinely amazed that such a vast criminal ecosystem has grown fat on unmissed scraps from programmatic advertising budgets. It might seem like a perfect crime, but that would be to ignore not only the cumulative waste of advertisers’ money but also the damage done to society by large-scale IP fraud of this type.
IP underpins the economy, and illegal streams of Premier League games, pay-per-view fights, films, subscription TV series and other commercial content not only hit the corporate profits of major rights owners and their legitimate distribution partners, but also diminish ongoing innovation, grass-roots funding and ordinary salaries up and down the content supply chain.
Now we are standing at a tipping point. Recently, the fightback has adopted meaningful new methods and recruited some important new friends. Substantial partners, including the large advertising exchanges, have joined the effort to defund pirate content, alongside numerous rights owners, regulators and law enforcement bodies.
And alliances with data partners and contextual engines now mean that advertisers can eliminate IP-infringing sites from their digital media schedules at the touch of a button. The automated, real-time scanning of the web and app landscape means that streams of live events can be detected within seconds, ads pulled, search rankings cut dead and sites taken down.
What was once a shadow industry about which brands and content owners knew little, and against which they had extremely limited sanctions, is now, thanks to the picture AI technology has been able to build, both highly quantifiable and eminently defundable. Now that we can see it, audit it and act upon it, we can demonstrate the nature of its threat to everyone who stands to lose from its success – and with their help, we can fight it.
[To share your insights with us, please write to firstname.lastname@example.org]