Spectrum of IP protection meets AI as a hybrid asset by Jutta Suksi

General context:

The visualization identifies:

• components used and generated during the machine learning process
• actions needed to be taken in each step of said process (horizontal axis)

and analyses how these components are situated in the spectrum of IP (intellectual property) protection, e.g. copyright and trade secrets (vertical axis).

Dimensions of the visualization in more detail:

The spectrum of IP protection has been identified on the left side of the visualization as the vertical axis. It identifies five clear categories of IPR (intellectual property rights), i.e. patent, trademark, trade secret, European database right and copyright; and additionally several other relevant areas of possible IP protection that does not account directly as IPR.

The machine learning process, on the horizontal axis identifies seven actions in the process (Build the untrained model, Get the training data, Train the model, Trained model, Apply trained model to data, Output, Use the output) with several steps within the actions.

Finally, the components identified with regard to each action have been plotted to the spectrum of IP protection.

Aims of the visualisation:

The illustration aims to gap knowledge from two very different areas, legal and technical, in order to understand the interlinkages, challenges and opportunities embedded in the use of AI in the society. Due to the vastness of the question, several simplifications to the details needed to be made, both in legal and in technological sense. The focus was given to the holistic understanding of the interlinkages.

Use of the visualization:

The visualization can be used as a map to identify core elements relating to different aspects around the generation of AI. With this map you can address the questions relating to AI in a more detailed manner, for instance, what kind of legal challenges arise with respect to different components involved or what kind of ethical questions may arise in different phases of the machine learning process.

Related information:

Visualization has been made in the project ZOOOM – ZOOOM – 3Os and IP awareness raising for collaborative ecosystems (zooom4u.eu) – funded by the European union.