Humanitarian AI Today Interview Series on Intellectual Property Law

Humanitarian AI Today is doing some podcast interviews on Intellectual Property Law, here’s some context and questions we’re interested in exploring.

At Humanitarian AI, a meetup.com community with local groups in fifteen cities, we’ve started work on setting up an open source foundation modeled on the Linux and Python Foundations to help keep algorithms, training and testing datasets and other utilities created by members free and accessible to others.

We’ve been thinking as such a lot about intellectual property law. For the benefit of members and the humanitarian community we thought we’d use the opportunity to do some Humanitarian AI Today podcast interviews on the subject of intellectual property law in the age of AI and explore IP law relevant to humanitarian applications of artificial intelligence.

Here’s some context and questions we’re interested in exploring.

Context

Outside of view of the tech community, humanitarian organizations have been working concertedly to make aid activities more transparent and to channel information on operations, transactions and results through open data sharing frameworks like IATI, managed by the International Aid Transparency Initiative that’s supported by the United Nations and mandated many government development agencies.

At the same time initiatives like the Center for Humanitarian Data are working on “increasing the use and impact of data in the humanitarian sector.” The Center’s vision is to “create a future where all people involved in a humanitarian situation have access to the data they need, when and how they need it, to make responsible and informed decisions.”

Humanitarian organizations are also openly storing information in centralized repositories like the Humanitarian Data Exchange or HDX and using hashtag schemes like the Humanitarian Exchange Language to make data comparable across organizational databases.

These initiatives are increasing the amount of data accessible to AI developers to aggregate and use to train, test and improve humanitarian AI applications. Soon, thanks to backend development work carried out by Humanitarian AI members and others, popular digital assistants like Alexa, Cortana or Siri will be able to plug into and use open data streaming through frameworks like IATI to answer complex queries posted by humanitarian actors, donor and others

But, on the flip side, humanitarian organizations are under pressure to reconsider the types of data they collect, be more careful about making data public and do more to protect the privacy and security of aid beneficiaries. Likewise on an operational level humanitarian organizations are increasingly being forced to be careful about information on aid activities being used by bad actors to identify and target hospitals and health clinics, aid convoys, refugee camps, aid access points and etc. These are very serious concerns in the field.

As AI developers aggregate humanitarian data and build applications capable of exposing information on aid activities and beneficiaries, it’s important for developers working on humanitarian open source initiatives for example to understand AI IP law, consider data ownership, data sharing implications and what being a responsible developer means in the age of AI?

Questions

We thought we’d ask our podcast interviewees to share their thoughts on questions like these:

• What’s artificial intelligence from an Intellectual Property Law vantage point? What does AI encompass: algorithms, machine learning models, training and testing datasets, data processing methodologies, intelligence, outputs and etc?
• What from an IP law vantage point is protectable?
• What consensus is taking shape around AI IP law and where does this consensus differ and by who?
• How is AI IP law being employed by startups, let’s hear some examples?
• How does IP law treat aggregated data published by others, used in AI training and testing datasets? How can this conceivably apply to data published by humanitarian organizations or by third parties like the Humanitarian Data Exchange or IATI?
• How do current and emerging open source licenses treat AI?
• Open source licenses and open source foundations help protect code committers from legal liabilities, what does this mean?
• How can open source initiatives potentially leverage IP law and open source licenses to help protect humanitarian actors and aid beneficiaries? Can some use cases come to mind?
• Should open source initiatives take responsibility for helping to limit the amount or types of information humanitarian algorithms can return? How could protecting intellectual property impact and slow or advance humanitarian AI development?

We’re not expecting any one person to address all these questions but we’re looking forward to hearing thoughts on the subjects from a range of interviewees.