NBC Universal patents system that tracks peer-to-peer swarms to catch online pirate seeders

Media conglomerate NBC Universal has patented a new system that uses artificial intelligence and big data to track peak periods when internet users are all illegally downloading the same specific file in order to figure out who the key perpetrators are.

The patented anti-piracy solution, entitled "Early Detection of High Volume Peer-to-Peer Swarms", was originally filed by NBC Universal in 2009, but was only finally approved by the US Patent and Trademark Office on 5 July 2016.

Many users around the world access and distribute pirated content over peer-to-peer networks using BitTorrent clients, where the file they want to download is distributed in segments called pieces to users, who are known as "peers". Once a user receives every single piece of the file, they can continue to share their copy with other users as "seeders".

The solution conceptualised by NBC Universal essentially lies in wait until users flock online to download a hugely popular file, such as a leaked copy of an unreleased movie or perhaps the latest episode of the hugely popular HBO series Game of Thrones.

The users access magnet links on online piracy sites, which open up in their BitTorrent clients and add them to a "swarm" where, as peers, their computers download segments of the file until it is completely downloaded.

Crawler bots infiltrate swarm to gain data on seeders

The swarm consists of both peers and seeders, and the largest swarms usually have the largest numbers of seeders. The idea is that the system will quietly send crawler bots into high volume swarms in order to gain information about only the seeders, such as their IP address, and the time and date that the torrent was being shared.

At the same time, the system also acts as a peer to download the same file that the rest of the users in each swarm is downloading, in order autonomously work out whether the file being downloaded is the same targeted file that the copyright holders are worried about.

If it is, in theory the system would be able to send alerts to internet service providers (ISP) as well as to tag "high-risk swarms" so that copyright holders are aware of them.

The solution is interesting, as in theory it could provide copyright holders with more real-time data on who exactly is downloading illegal content and sharing it online, and when online pirates are most active. However, in practice, the solution would still require a certain amount of time to analyse swarms, such as a period of up to 48 hours, as used in an example cited in the patent filing.

But will this solution actually stop online piracy?

The thing is, after a file is released onto the internet – such as the latest instalment of Game of Thrones after it airs on Sunday nights – the amount of users downloading the file is at its peak usually within the first 24-48 hours. This is when the swarms are at their largest.

Also, with today's internet speeds, it takes users roughly about 20 minutes at most to download an episode of Game of Thrones, and many users prefer to download a file and then go offline without sharing the file with anyone, which means that they stay as peers.

This means that after 20 minutes, when the file has completely downloaded, they essentially leave the swarm, so the swarm will get smaller. Of course, NBC Universal doesn't really care about peers – they want to zero in on the most prolific seeders and find a way to bring them down. But their solution doesn't describe how you do this.

The copyright holders would still need to contact the user's ISP demanding that the ISP identify the customer by their IP address, and if the ISP said no, the copyright holders would still need to take the ISP to court and subpoena the information, before taking the identified user to court for copyright infringement. All of these things take months, by which time, the files would already have been shared millions of times and users might not be downloading them any more.