Applications closed

The Online Safety Benchmark request for proposals


The reduction of fake and misleading content on Facebook is mostly driven by the state-of-the-art text and visual recognition systems, including Machine Translation, Automatic Speech and Character Recognition, and Image and Text Categorization. However, we still need major improvements in AI systems to further improve online safety. What is currently lacking is a well-defined set of tasks around online safety, together with appropriate benchmarks to quantify performance.

The goal of the RFP is to challenge the community to address these challenges together, and to solicit new tasks.

To make progress on the major social issues in society, Facebook is partnering with universities to support them in building open-source datasets on which the community can measure the progress of existing techniques to reduce misleading behavior online.

Applications Are Currently CLosed

Application Timeline

Notifications will be sent by email to selected applicants by mid-July.

Launch Date

May 6, 2019


May 30, 2019

Winners Announced

Mid-July 2019

Areas of Interest

The goal of this RFP is to help the academic community to address problems in the area of safer online conversations. This includes problems around misinformation, as well as hate speech and inauthentic online behavior, to name a few. The grant aims to provide funding for projects that build research infrastructure such as datasets or evaluation platforms that can accelerate research in a broader way. In particular, we are encouraging:

  • The creation of a publicly available benchmark and analysis platform that enables the community to compare and understand models that address problems in the area of safer online conversations. An example of an existing platform (but outside that area) is General Language Understanding Evaluation, a.k.a. GLUE, which subsumes existing NLP datasets, provides online evaluation tools and leaderboards, and specific diagnostic evaluation data to test specific system properties.
  • The creation of publicly available datasets to improve the state-of-the-art of models for safer online conversations. Examples are fake news datasets, such as BuzzFeedNews or LIAR, rumor propagation datasets such as the Kaggle Rumor Tracker Dataset or offending content datasets like the MS Offensive Language Dataset.

The funding can range from $10K to $50K, depending on the proposal, but should roughly match the cost of annotation, e.g. using a crowdsourcing annotation platform or paying expert annotators.


  • Privacy: Unless there is an explicit user consent to share their data, user-specific information should be properly anonymized. The method used to anonymize the data must be clearly explained.
  • Data license. The recipient of the grant will agree to open source the dataset to the research community. The CC Attribution license is recommended, but alternative license agreement (e.g. for labelled content under existing copyright) can be considered if the proposed license allows access, processing and storage for non-commercial use.
  • Language. There is no language restriction for the annotation, although English translation is recommended for easier reproducibility.


Proposals should include

  • A summary of the project (1-2 pages) explaining the area of focus, a description of techniques, any relevant prior work, and a timeline with milestones and expected outcomes.
  • A draft budget description (1 page) including an approximate cost of the award and explanation of how funds would be spent. A separate line for travel expenses to the Truth and Trust Online conference in London, UK should be included.
  • Curriculum Vitae for all project participants.
  • Organization details. This includes tax information and administrative contact details.


  • Awards must comply with applicable US and international laws, regulations, and policies.
  • Applicants must be current full-time faculty at an accredited academic institution that awards research degrees to PhD students.
  • Applicants must be the Principal Investigator on any resulting award.
  • Applicants may submit one proposal per applicant.
  • Organizations must be a nonprofit or non-governmental organization with recognized legal status in their respective country (equal to 501(c)(3) status under the United States Internal Revenue Code).

Additional Information

For questions related to this RFP, please email

Terms & Conditions

  • By submitting a proposal, you are authorizing Facebook to evaluate the proposal for a potential award, and you agree to the terms herein.
  • You agree that Facebook will not be required to treat any part of the proposal as confidential or protected by copyright.
  • You agree and acknowledge that personal data submitted with the proposal, including name, mailing address, phone number, and email address of you and other named researchers in the proposal may be collected, processed, stored and otherwise used by Facebook for the purposes of administering the website and evaluating the contents of the proposal.
  • You acknowledge that neither party is obligated to enter into any business transaction as a result of the proposal submission, Facebook is under no obligation to review or consider the proposal, and neither party acquires any intellectual property rights as a result of submitting the proposal.
  • Any feedback you provide to Facebook in the proposal regarding its products or services will not be treated as confidential or protected by copyright, and Facebook is free to use such feedback on an unrestricted basis with no compensation to you.