Image by starline on Freepik
EUIPO’s Study on the Impact of Artificial Intelligence on the Infringement Enforcement of Copyright and Design provides some great examples of AI application to protect the Intellectual Property rights by the law enforcement, that could be used in the private sectors.
– Physical product copyright and designs:
- Theft of copyrightable work or design under development: theoretically, machine supervised learning can be used to break the encryption of organised criminal groups (OMD). Moreover, convolution neural networks (CNN) might be utilised to classify encryption schemes. Natural language processing (NLP) – enabled filters may classify and analyse emails to block phishing attacks. And, AI bots can detect targeted social engineering bots.
- Mass production of copyright and design-infringement: the CNN can analyse and recognise visual imagery to scan shapes and patterns for similarities to detect the infringement.
- Online marketing of copyright and design-infringing products: web crawlers can find patterns (e.g. domain names, content of websites with fakes, websites or products with similar names, or Javascript) of the OMDs. NLP can also analyse users’ behaviour and their content shared in many languages and track the marketing of the infringing products by identifying the characters and similarity to the originals.
– Copyright and designs of digital content
- Hacking media accounts: CNN for image and pattern recognition can identify online attempts at social engineering. Machine learning (ML) also can identify fake login pages, phishing behaviours and block these suspicious activities in addition to computer vision and NLP.
- Social media offences: Pattern recognition can identify social engineering risk and AI bots, such as automated cyber patrols, can look for images on social media, containing copyright and design infringement.
- Unauthorised streaming of broadcasts: AI-based web crawlers can find the patterns to identify criminal actors, techniques, trends. ML can identify copyright infringement by analysing the sequence of images in videos. NLP can find correlations in datasets on a website or exploratory analysis, etc.