Facebook Ethics in AI Research Initiative for Africa 2020
Application Deadline: 22nd July 2020 8:00 a.m. PST (5:00 p.m. CAT)
About the Award: This research initiative complements Facebook’s existing efforts to bolster independent research being undertaken in these areas, such as those we are supporting in the Asia Pacific, India, and Latin America, and at the TUM Institute for Ethics in AI.
This research initiative, through a request for proposals (RFP) coordinated by the Human Sciences Research Council (HSRC) of South Africa, will result in the production of papers from academic institutions, think tanks, and research organizations registered and operational across Africa. Proposals will be reviewed by a selection committee, and the entities whose proposals are selected will receive a research grant.
Eligible Field(s): We are particularly interested in proposals that pertain to the following research topics and questions. However, these are just examples of possible research questions. Applicants are encouraged to come up with their own proposals and to leverage, as much as possible, regional context and local examples in their submissions.
Ethics by design
- Policymakers are emphasizing the need for ethics by design. What does ethics and AI mean and look like in an African context? How can academia, government, and industry collaborate to promote and advance ethics by design practices and frameworks?
- How can developers and companies ensure that their AI systems are explainable, what their purpose is, and what they entail?
- How can academia help companies and governments better understand and operationalize ethics within their own sectors and activities? How can academia both inspire and build on industry best practices for the responsible and ethical development of AI?
- How can developers and companies ensure that their AI systems and applications are built in a fair and unbiased way? How should social science and humanity’s questions around fairness and discrimination be embedded in the technical design of AI? What best practices can we advance in this space?
- How can developers and companies ensure that their AI systems are transparent to the affected individuals in a meaningful way? (For example, how will a person know whether they are being unfairly discriminated against because of an automated decision?)
- What level of autonomy should be provided to the affected individuals concerning the use of AI without compromising the legitimate purposes of AI, and how should individuals exercise control over automated decisions?
Governance, ethics, and human rights
- How do hard and binding legal instruments, namely existing and proposed legislation, and ethical AI governance frameworks (soft law, non-binding) interact and mutually influence each other? Does the former stem from or pre-empt the latter? Are they complementary or competing?
- How should international, regional, and national human rights frameworks coexist and interact with each other? What can AI and ethics learn from human rights law and from the obligations it places on public and private actors?
- What is the role of ethical codes in the broader regulatory landscape? How should they relate to laws and regulations (either existing or being developed)?
- What is the role of academia in 1) promoting research on AI governance frameworks; 2) analyzing, anticipating, and identifying gaps in legislation and other governance models related to AI development and use; and 3) articulating best practices to guide ethical and innovative uses of data?
AI ethics and diversity
- What factors should inform human and technical diversity in the design and development of AI systems (e.g. linguistic diversity, diversity of race, gender, religion, and so on) in the African context?
- How should AI developers and companies foster and apply a multicultural approach to the ethical design of AI?
- What are the most prominent narratives on the role and impact of AI in Africa? How does local knowledge on the relationship between humans and machines shape the understanding, the perception, and adoption of AI in Africa?
- How should companies navigate the tension between the benefits of a global approach versus the need to acknowledge important particularities and differences stemming from companies’ regional user base?
- What would be a sound and scalable methodology for researchers and product developers to recognize and solve ethical challenges, while leveraging different regional perspectives that promote diversity?
Eligibility of the Facebook Ethics in AI Research:
- Applying organizations must be nonprofit or nongovernmental organizations with recognized legal status in their respective country (equal to 501(c)(3) status under the United States Internal Revenue Code) and should comply with the applicable laws to receive foreign funds in their respective country.
- Applicants must be the principal investigator (PI) on any resulting award.
- Researchers are also responsible for obtaining approval from the PI’s university’s ethics/institutional review board, if applicable.
- Each applicant may submit only one proposal per solicitation.
Proposals should include
Proposals should be 2–3 pages and include the following:
- A research question and a clear statement of work that can be completed within six months of the award
- A summary of the project (1–2 pages) explaining the area of focus, a description of techniques, relevant prior work, and a timeline with milestones and expected outcomes (note that the timeline should not be more than six months)
- A draft budget description (<1 page) including an approximate cost of the award and explanation of how funds would be spent. Proposals are encouraged to focus on funding on project personnel, especially Ph.D. students. Proposals from small collaborative teams are also encouraged. The proposed budget should be within $25,000 (the overhead is typically limited to 5–8%). Please note that in the submission form, you will be asked to provide the requested amount in U.S. dollars.
- Name(s) of the researcher(s) involved in the proposed work with their CV/resume
- Indication of any previous or current connections/collaborations with Facebook (in which case, please provide context and name the Facebook contacts)
To be Taken at (Country): USA (tentative)
Number of Awards: Not specified
Value of Award:
- The proposed budget should be within $25,000 (the overhead is typically limited to 5–8%). Please note that in the submission form, you will be asked to provide the requested amount in U.S. dollars.
- If Facebook is offering event attendance and related travel to grant recipients, the specific travel benefits, travel dates, and additional details (collectively, “Travel Benefits”) will be provided by Facebook to the grant recipient.
How to Apply for the Facebook Ethics in AI Research: