Legal advice for software, AI, and data-driven businesses.
Aligning commercial, technical and legal to address the unique challenges and opportunities in this space.
Delivered by a single multi-disciplinary team.
We offer joined-up thinking across a comprehensive range of services:
Paths through the neural network represent how certain features of data and algorithms can engage different legal considerations in the Legal Layer. Our Codiphy team considers these all together to optimise the outcomes for your business.
Click the buttons to show some examples of the many paths through the network.
Some AI models are trained on data scraped from publicly-available sources. This may have legal implications. For example, training an AI model using copyright works without obtaining a licence from the rightsholder could result in litigation and/or reputational damage. AI regulations will also play a role. For example, the EU AI Act will oblige General Purpose AI providers to publish summaries of training data and to respect opt-outs from data scraping. Otherwise, they could face penalties under the Act.
Proprietary algorithms may involve IP that can be protected, for example using trade secrets or patents. Competing factors need to be considered to optimise protection of that IP. For example, trade secret protection may extend an early-mover advantage, but will be lost if technical information is publicly leaked. Patents can provide protection even if the technical information is later published or if someone else develops the same idea independently, but filing a patent application for an idea that isn’t patentable may result in a loss of IP. An effective IP strategy will consider these factors and others to maximise the value in the IP.
Original code is protected by copyright, which may be licensed for others to use, for example under an open-source licence. Non-compliance with the obligations of an open-source licence may result in litigation and/or reputational damage. If you are considering releasing code under an open-source licence, it is important that this decision fits in with your business model and wider IP strategy, and that the licence you use supports these. Attempting to change licence terms after release can be difficult or impossible, and can cause reputational damage in the eyes of the open-source community.
AI models may be trained on, or may subsequently process, personal data that relates to an identified or identifiable natural person (including personal data that is publicly accessible). This may have legal implications if the training/processing is not compliant with data protection obligations. The party controlling the processing could face regulatory penalties including orders to stop the processing or using the AI model, litigation between the user and the controller, and/or reputational damage. Providers of AI systems should consider the principle of ‘privacy by design’, as well as ensuring contractual provisions are in place to mitigate the legal risk.
Our advisors bring together a wealth of experience in digital IP protection, software development, and commercial law. We understand the importance of both safeguarding your revenue streams and investment through IP rights, and promoting collaboration to help you achieve your objectives.
Be kept up to date with all things Codiphy at EIP and from around the world. Listen to our EIP talks podcast, read our insights and published publications and more.
+44 (0)20 7440 9510
codiphy@eip.com
or Enquire now
London +44
(0)20 7440 9510
