Case Studies: Choosing an ML License#

Below are two hypothetical case studies based on concerns raised through the license selection process during BigScience, a one-year long research workshop on large multilingual models and datasets.

Responsible AI License for Research & Academia#

Cidney, a hypothetical ML researcher, is working for a university lab, has developed as part of her PhD research an ML vision model for facial recognition. She is well aware of both the good quality of the model and its limitations, and is willing to inform users about it in the model card she is planing to release along with the model.

Cidney really wants to openly release the model to foster further research in the field, for other researchers to test it and provide feedback to her, or even to come up with improved versions of it.

However, she is concerned about potential uses of the model which might lead to undesired outcomes, according to what she thinks is not good to use the model for, and also informed about her research lab ethical guidelines and code of conduct.

Consequently, she decides to use a Responsible AI License to release her ML Model:

  1. She wants to place use-based restrictions for specific identified scenarios informed by her research literature review, her experience, and her awareness of the model technical limitations. Thus, she comes up with a set of scenarios where due to her concerns and the technical capabilities of the model she does not feel comfortable that the model is used for. Eventually, when drafting the restrictions, she will ask for legal advice from the university legal staff.

  2. She decides to license just the pre-trained model, not the training dataset nor an app embedding it. The source code is already available out there with an open source license, so there is no need to license it again.

  3. Her research lab allows her to release the model on an open basis for research purpose, thus enabling free access and distribution of the model solely for research and non-commercial purposes.

  4. As a result, Cidney will use a RAIL-M license to release her model. Even though the model is accessible on an open basis, the license does not allow for a permissive distribution of it for commercial purposes.

Responsible AI License for Industry#

HealthyCo is a hypothetical ML startup focused on the health sector and developing innovative solutions for market niches such as medicines testing processes, cardiovascular predictive algorithms, and protein folding, among other areas.

HealthyCo is working on a platform wherein companies will be able to integrate their pipelines and end-consumers will be able to share the data generated by their smart-watches when running. The company is aware of the fierce competition in the market and wants to leverage the network effects-related capabilities of an open platform. Therefore, HealthyCo is considering to release several of its ML models, but also fine tuned versions embedded in specific software apps. The goal is to generate traction and foster adoption of the platform in the short term while allowing companies and users to test and experience. Moreover, they are already in talks with some investors and have agreed to start discussions for their next investment round in 3 months.

HealthyCo is well aware of the capabilities of their ML models and apps, and according to its values as a company striving for ethics-informed research and ML development, it is reluctant to place the technology in the market under an open source license, as users could do whatever they would like with the technology and the company would not be able to control potential misuses.

Accordingly, the startup decides to release the ML models and apps with Open & Responsible AI licenses. On the one hand, they would like companies with whom they are collaborating and other researchers to access, test and further improve or build upon their products. On the other hand, they know investors would highly appreciate that the company is taking pioneering steps forward towards a responsible use and distribution of ML artifacts in the AI space, as a much needed and new trend.

HealthyCo decides to use an Open & Responsible AI License for its cardiovascular prediction ML app:

  1. It wants to place use-based restrictions for specific identified scenarios informed by the company researchers’ experiments and findings, and its awareness of the ML app capabilities. Thus, HealthyCo comes up with a set of scenarios where due to its concerns and the technical capabilities of the model the startup does not feel comfortable that the model is used for.

  2. HealthyCo decides to license a fine tuned model embedded in a software app, but not the training dataset. The use-based restrictions will apply to both the ML model and app.

  3. It seeks to release the model on an open basis enabling flexible downstream distribution for commercial purposes also.

  4. As a result, HealthyCo will use an OpenRAIL-AM license.