With the launch of Stanford’s Human-Centered AI Institute (HAI), we are seeing enthusiasm across the University and its disciplines for the Institute and its tripartite mission of basic AI research, applications across diverse areas, and understanding the human and societal impacts of the technology. A common question from students and faculty is “How can I get involved in HAI?” Although there are some administrative leaders of the effort, it is clear that in order to have the most impact across HAI’s mission areas we will require the effort and creativity of our community.
In general, a new institute or center brings people together in three ways:
-
Multi- and interdisciplinary research projects
-
Training opportunities (at all levels)
-
Sharing research infrastructure
These are the primary places for engagement with HAI and offer these thoughts about moving new ideas forward.
- Multi- and interdisciplinary research projects
HAI already has a seed grant program which has created some amazing new projects. We are in planning phases for additional projects that offer mid-level and high-level funding. The most compelling projects seem likely to address more than one of the mission areas, because the interstices are where many of the challenges remain. Thus, while purely disciplinary research questions are not out of bounds, they may sometimes be less attractive than ones that bring together scholars who might otherwise not work together. In addition, the most compelling proposals (especially seed grants) will have some articulated path for follow-on funding outside the HAI. HAI is not primarily a funding agency, and so research funding is almost always meant to begin and stimulate efforts that will find sources of support outside the Institute. This goal brings another research-related opportunity: attracting industry (and other) independent collaborators and funding into HAI. A key part of the HAI mission is to impact the world through collaborative work with those bringing AI systems to scale, and so leadership in attracting and maintaining relationships with external research and development partners is particularly important.
Sometimes a group may feel that there may be a great opportunity for collaborative research, but the projects and goals are not yet obvious. They may simply need time to just get to know one another in an informal or semi-formal setting to understand how the problem spaces they are addressing and current/anticipated AI capabilities might impact those spaces. HAI is open to compelling proposals to catalyze these informal interactions so that they can gain focus and structure for more formal proposals.
- Training opportunities
Critical to the success of HAI is building an ecosystem of education for undergraduates, graduate students, post-doctoral fellows, staff and faculty. In addition, our outreach mission makes professional education for partners in industry, non-governmental organizations, and government absolutely critical as well. We recognize the vast range of learning opportunities that are required to prepare leaders for the opportunities and challenges of deploying AI systems in their fields of work—ranging from very basic introduction to high level ideas and capabilities, to in-depth transfer of technical knowledge about systems design, building, and validation. Thus, we need the ideas and energy of participants to identify specific teaching/learning opportunities and helping organize these in terms of target audience, curriculum, evaluation, and learning objectives. This is an exciting way to build relationships that may then spill over into research and social impacts work, so (as usual!) the training mission is the straw that stirs the entire Institute.
Another important element of training is bringing people on campus to enrich our environment for short/medium/long term stays as fellows. What can these people offer? How will their stay be structured? How will the HAI community benefit from their presence and activities, and how will they bring what they learn back to impact the world more broadly?
- Shared research infrastructure
Infrastructure to support research takes many forms. The most basic and obvious might be compute infrastructure and hardware infrastructure. Other critical resources include important data sets—static as well as “online” piped real-time data. But research infrastructure can also take the form of strategic relationships with external entities, focus groups of excellence that create research task forces, communications expertise for writing and speaking about AI, HAI, and the big challenges to the field. The HAI leadership is depending on our colleagues at the University to identify specific opportunities, and take the lead on creating/shepherding these resources so that they can benefit the effort in tangible ways that advances our core agendas of research and teaching.
For all these activities, HAI has funds to support the most compelling proposals. We hope to see a bounty of good ideas, and have thought carefully about the importance of clear criteria for supporting proposals for new initiatives. They will vary with particular programs, but we encourage participants to consider these broad categories:
-
How does the activity support the three core activities of the HAI? Does it bring challenges to one of the core areas, or more?
-
Is this a one-off event or is it a longer-term activity being launched?
-
Which Stanford scholarly communities and individuals (and communities/individuals outside Stanford) stand to benefit from this activity?
-
Is the activity designed to maximize inclusion and diversity, along all relevant dimensions, and “expand the umbrella” of those working on AI and its promise?
-
What is the required/requested budget? Have other sources been identified as a source of matching/replacing HAI funding? Is there a business plan for the activity after an initial startup period?
What research disciplines will be engaged by the activity? Are they frequent fliers in the AI discussion or are new constituencies engaged?
In the end, the Directors and Associate Directors of HAI are looking to “deputize” our colleagues to take responsibility and leadership in these new activities. These deputies will build up a resume of contributions and will naturally form the HAI community and help drive decisions and resource allocation for the Institute as it moves forward in the next decade.
We hope you will join us in building HAI. If you would like to make a suggestion or proposal, please send us an e-mail. We will consider all thoughtful proposals from faculty, fellows, or students of any school, department, institute, or center of Stanford at any time.
And don’t forget to sign up for our mailing list to receive the latest HAI news and updates, and to learn about the upcoming opportunities to participate in events, calls for proposals, or other activities.