AI and data science
Developments in AI, machine learning, data science and big data is driving business and innovation across many sectors – from speech recognition, computer vision, autonomous vehicles and document interpretation to social media, life sciences, healthcare, agriculture and numerous others. There are many intellectual property considerations that apply to data and data processing – from patents and trade secrets to copyright and database rights.
For many of our clients, data is a core component of their business. The value of data stems from the information it contains, and the ability to extract and harness that information commercially. This means there is value in the raw data, and also in the intellectual property rights associated with its processing, and the output of that processing.
Innovation can arise from technical solutions at any stage of the data processing pipeline. We have helped our clients to identify and protect innovations around:
- data science, data mining and data analytics;
- big data;
- hardware for managing massive data requirements, sharding and server architectures;
- machine learning and data-driven models;
- combinations or applications of existing machine learning models, including open source;
- feature learning or extraction;
- training, transfer learning (pre-training or fine-tuning) and the creation or labelling of training data;
- data cleaning;
- user interfaces; and
- data structures and databases.
We can help you protect such innovations with patents, identifying circumstances in which patent protection would be appropriate and will add value to your business.
In certain circumstances, you may be able to rely on the legal provisions around confidential information or trade secrets. This can give a commercial lead, but it is important to understand the limitations of these provisions. They do not always apply, and rights to confidential information or trade secrets can easily be lost if they are not managed properly. We can help you to determine when this strategy is appropriate and guide you through the steps needed to maintain such rights.
The law also provides copyright protection and sui generis protection for databases in certain circumstances, which may apply to the original data and to the output of the data processing or model weights learned from it.
We also understand the role of scientific publications and open source in the AI and data science communities and can help you manage this effectively in the context of your business's intellectual property strategy.
Artificial intelligence also raises the intriguing question of how to handle the IP in an innovation which arises as a result of AI programs.
In 2020, two of our leading experts in AI, Virginia Driver and Tom Woodhouse, contributed to the intellectual property chapter of The Law of Artificial Intelligence, published by Sweet and Maxwell.
Click here for the contents of the chapter on IP, which covers:
- patentability of AI, AI-devised or AI-assisted inventions;
- inventiveness in using AI;
- copyright and weaknesses in protection;
- data and database rights;
- design rights – registered and unregistered; and
- trade marks.