Yesterday, the US and UK governments announced the 12 winners of the first phase of the U.K.-U.S. privacy-enhancing technologies (PETs) prize challenges.
Competing for cash prizes from a combined U.S.-U.K. prize pool of $1.6 million (£1.3 million), participants are developing solutions that will enable artificial intelligence models to be trained on sensitive data without organizations having to reveal, share or combine their raw data.
Winning challenge solutions will be showcased at the second Summit for Democracy, which President Joe Biden plans to convene in the first half of 2023.
The U.S. challenge is funded and administered by the U.S. Department of Commerce’s National Institute of Standards and Technology (NIST) and the U.S. National Science Foundation (NSF).
“AI is driving rapid technology change that is based on ever-increasing amounts of disparate data, making privacy-enhancing technologies increasingly important,” said Under Secretary of Commerce for Standards and Technology and NIST Director Laurie E. Locascio. “The U.S.-U.K. PET prize challenge provides a global venue to build and showcase cutting-edge and scalable solutions that respect human rights and civil liberties. I am excited by the solutions proposed by these scholars and look forward to their impact on enhancing privacy and bolstering trustworthy AI.”
“These first-of-their-kind international prize challenges are focusing innovators from the U.S. and U.K. on overcoming the challenge of maturing PETs for practical use cases,” said National Science Foundation Director Sethuraman Panchanathan. “The level of participation and caliber of participants in the U.S.-U.K. PETs prize challenges promise to accelerate the translation of PETs to practice. I’m excited to see the strong start to the prize challenges across the transatlantic research community and look forward to the results in the next phases.”
The 12 prize-winning technical papers were selected from 76 entries across two challenge tracks: using PETs to improve detection of financial crime and forecasting an individual’s risk of infection during a pandemic, or designing a solution that would meet both scenarios. A total of $157,000 (£138,000) was awarded in this phase.
“Privacy-enhancing technologies have the potential to unlock the power of data to tackle major societal challenges — from international money laundering to responding to global pandemics — in a way that respects citizens’ rights,” said Minister of State for Media, Data, and Digital Infrastructure Julia Lopez at the UK Department for Digital, Culture, Media and Sport. “That’s why I’m delighted by the strength of the response to the joint U.K.-U.S. prize challenges, with world-class researchers on both sides of the Atlantic leaping to the challenge of innovating in a way that upholds our shared values.”
For the first stage of the competition, teams from academic institutions, global technology companies and privacy tech companies submitted white papers describing their approaches to privacy-preserving data analytics. In the second phase, which began in October 2022, the teams began building the solutions they set out in their white papers. The challenge participants will have opportunities to engage with regulators and government agencies for guidance on developing solutions that uphold crucial regulatory principles.
In the U.S., prizes were awarded to the following teams:
- Team MusCAT — researchers from the Broad Institute, Massachusetts Institute of Technology, Harvard Business School, University of Texas Austin, University of Toronto
- Team IBM Research
- Team Secret Computers — researchers from Inpher Inc.
In the U.K., prizes were awarded to the following organizations:
- Corvus Research Limited
- Diagonal Works
- Featurespace Limited
- OpenMined and DeepMind
- Privitar Limited
- University of Cambridge
- University of Liverpool
In the third phase of the prize challenge, “red teams” will launch privacy attacks on the most promising solutions to rigorously test their privacy-preserving capacities and assess the final winners. Applications for red teams are being accepted through Nov. 23, 2022. Top-scoring red teams will be awarded prizes from a combined pool of $225,000 (£200,000).
Solutions for the financial crime and public health tracks are being developed using synthetic datasets that are representative of real-world use cases. The prize challenge organizers have announced today that the synthetic banking data used in the financial crime challenge is based on data by global financial institutions BNY Mellon and Deutsche Bank.
Innovators on the financial crime track are also working with synthetic global transaction data created by SWIFT, a global provider of secure financial messaging services, with the support of MOSTLY AI. Innovators on the public health track are working with a synthetic dataset created by the University of Virginia’s Biocomplexity Institute.
Planning for the challenges is being led by the UK Centre for Data Ethics and Innovation and Innovate UK, and the U.S. White House Office of Science and Technology Policy, NIST and NSF.