IP-AI •
What does the future look like for AI?
Dr. ʻŌiwi Parker Jones (Kanaka Maoli) is a Research Fellow at the University of Oxford where he works on biological and artificial intelligence in the departments of Neuroscience and Engineering. He attended the March 2019 Indigenous Protocol and Artificial Intelligence workshops in Hawai’i. Here he explores the future of AI.
My first thought on the topic was to frame it in terms of AI by Indigenous communities and for Indigenous communities. But I have also, more recently, been considering a third way: AI in dialogue with Indigenous communities.
I would propose the following working definitions for the three views:
AI by an Indigenous community is AI that is produced by one or more members of an Indigenous community.
AI for an Indigenous community is AI that addresses the needs of one or more Indigenous communities.
AI with an Indigenous community is AI that is in dialogue with one or more Indigenous communities.
Here (1) is intended to denote anything produced by a member of an Indigenous community, no matter what. So if a member of an Indigenous community worked on any random topic in machine learning, then, by definition, it would be ‘Indigenous AI’. To me this misses the point. As an Indigenous person who works on AI, I appreciate the sentiment. But if I invented a new kind of LSTM module, should that module be considered ‘Indigenous AI’? We could end up with an incoherent subset of AI research that we call ‘Indigenous AI’ simply because Indigenous people worked on those things.
Definition (2) focuses on the content of the research, rather than on the identity of the researcher. AI for an Indigenous community might include some of my own work on Hawaiian NLP. Should any research that touches on topics relevant to an Indigenous community be considered ‘Indigenous AI’? One limitation of (2) is that it does not give agency to our Indigenous communities over what counts as ‘Indigenous AI’. Any company might, for example, develop an application for one of our languages, or for any part of our culture, and then market it as ‘Indigenous AI’. Is that the space that we want to create around this term?
Definition (3) is meant to maximise the pros and minimise the cons of (1) and (2). ‘In dialogue with’ is meant to express the idea that the AI is being actively engaged with by an Indigenous community. One reason to engage with AI research might be that it is being performed by someone who is already a member of the community, as in (1). Another reason is that the AI research bears on topics that are important to the community, as in (2). But definition (3) leaves the choice about what counts as ‘Indigenous AI’ up to our communities, so that that it should be impossible to hijack the term without buy-in from at least one of our communities.
This, I would suggest, is one way to frame what we will be doing at the workshop: entering into dialogue between research on AI and our Indigenous communities.
From this perspective, then, what does the future of Indigenous AI look like? This question has been posed by the workshop organisers. If I could suggest a few relevant topics, they would include: intellectual property, fairness, and data-efficiency. I hope that we will get to talk more about these things at the workshop. However, if the big idea is to create a community of ideas, then I also look forward to finding out what ‘Indigenous AI’ means together.
I also hope that we might continue to broaden our conversation to include more non-Indigenous AI researchers, with the intention of producing as active an ecosystem of ideas together as we can.
Author Bio:
Dr. ʻŌiwi Parker Jones (Kanaka Maoli) is a Research Fellow at the University of Oxford where he works on biological and artificial intelligence in the departments of Neuroscience and Engineering. In the 1980s, he was among the first children to be raised speaking Hawaiian in two generations. Later, as a graduate student, he worked on the adaptation of big data computing for the often fragmented corpora available in endangered languages—a research programme that he has continued to advance, for example by developing hybrid Deep Learning methods that contribute to the preservation and revitalisation of the Hawaiian language (e.g. Shillingford and Parker Jones 2018). As a postdoc, Dr. Parker Jones trained in systems neuroscience—with an emphasis on applications of machine learning to large-scale brain data. His current research is focused on Brain Computer Interfaces.
The Indigenous Protocols and Artificial Intelligence (IP-AI) workshops are founded by Old Ways, New, and the Initiative for Indigenous Futures. This work is funded by the Canadian Institute for Advanced Research (CIFAR), Old Ways, New, the Social Sciences and Humanities Research Council (SSHRC) and the Concordia University Research Chair in Computational Media and the Indigenous Future Imaginary.
IP-AI •
What does the future look like for AI?
Dr. ʻŌiwi Parker Jones (Kanaka Maoli) is a Research Fellow at the University of Oxford where he works on biological and artificial intelligence in the departments of Neuroscience and Engineering. He attended the March 2019 Indigenous Protocol and Artificial Intelligence workshops in Hawai’i. Here he explores the future of AI.
My first thought on the topic was to frame it in terms of AI by Indigenous communities and for Indigenous communities. But I have also, more recently, been considering a third way: AI in dialogue with Indigenous communities.
I would propose the following working definitions for the three views:
AI by an Indigenous community is AI that is produced by one or more members of an Indigenous community.
AI for an Indigenous community is AI that addresses the needs of one or more Indigenous communities.
AI with an Indigenous community is AI that is in dialogue with one or more Indigenous communities.
Here (1) is intended to denote anything produced by a member of an Indigenous community, no matter what. So if a member of an Indigenous community worked on any random topic in machine learning, then, by definition, it would be ‘Indigenous AI’. To me this misses the point. As an Indigenous person who works on AI, I appreciate the sentiment. But if I invented a new kind of LSTM module, should that module be considered ‘Indigenous AI’? We could end up with an incoherent subset of AI research that we call ‘Indigenous AI’ simply because Indigenous people worked on those things.
Definition (2) focuses on the content of the research, rather than on the identity of the researcher. AI for an Indigenous community might include some of my own work on Hawaiian NLP. Should any research that touches on topics relevant to an Indigenous community be considered ‘Indigenous AI’? One limitation of (2) is that it does not give agency to our Indigenous communities over what counts as ‘Indigenous AI’. Any company might, for example, develop an application for one of our languages, or for any part of our culture, and then market it as ‘Indigenous AI’. Is that the space that we want to create around this term?
Definition (3) is meant to maximise the pros and minimise the cons of (1) and (2). ‘In dialogue with’ is meant to express the idea that the AI is being actively engaged with by an Indigenous community. One reason to engage with AI research might be that it is being performed by someone who is already a member of the community, as in (1). Another reason is that the AI research bears on topics that are important to the community, as in (2). But definition (3) leaves the choice about what counts as ‘Indigenous AI’ up to our communities, so that that it should be impossible to hijack the term without buy-in from at least one of our communities.
This, I would suggest, is one way to frame what we will be doing at the workshop: entering into dialogue between research on AI and our Indigenous communities.
From this perspective, then, what does the future of Indigenous AI look like? This question has been posed by the workshop organisers. If I could suggest a few relevant topics, they would include: intellectual property, fairness, and data-efficiency. I hope that we will get to talk more about these things at the workshop. However, if the big idea is to create a community of ideas, then I also look forward to finding out what ‘Indigenous AI’ means together.
I also hope that we might continue to broaden our conversation to include more non-Indigenous AI researchers, with the intention of producing as active an ecosystem of ideas together as we can.
Author Bio:
Dr. ʻŌiwi Parker Jones (Kanaka Maoli) is a Research Fellow at the University of Oxford where he works on biological and artificial intelligence in the departments of Neuroscience and Engineering. In the 1980s, he was among the first children to be raised speaking Hawaiian in two generations. Later, as a graduate student, he worked on the adaptation of big data computing for the often fragmented corpora available in endangered languages—a research programme that he has continued to advance, for example by developing hybrid Deep Learning methods that contribute to the preservation and revitalisation of the Hawaiian language (e.g. Shillingford and Parker Jones 2018). As a postdoc, Dr. Parker Jones trained in systems neuroscience—with an emphasis on applications of machine learning to large-scale brain data. His current research is focused on Brain Computer Interfaces.
The Indigenous Protocols and Artificial Intelligence (IP-AI) workshops are founded by Old Ways, New, and the Initiative for Indigenous Futures. This work is funded by the Canadian Institute for Advanced Research (CIFAR), Old Ways, New, the Social Sciences and Humanities Research Council (SSHRC) and the Concordia University Research Chair in Computational Media and the Indigenous Future Imaginary.