Austrian synthetic data pioneer MOSTLY AI has launched a $100,000 global challenge to drive adoption of privacy-safe synthetic data and highlight its potential to safely fuel artificial intelligence innovation.
Dubbed The MOSTLY AI Prize, the challenge invites data scientists, AI developers, and researchers to create high-fidelity synthetic datasets from real-world data. Entries will be judged on accuracy, privacy, usability, and generalisability, with the aim of showcasing synthetic data’s role in powering safe, open-access AI.
The prize pool – the largest yet for a synthetic data challenge – is split between two tracks: the Flat Data Challenge, involving static, table-based data like patient records, and the Sequential Data Challenge, for time-ordered datasets such as stock values or longitudinal health data.
Entrants must submit anonymised synthetic datasets that closely mirror the original data while maintaining privacy and complying with regulations. Submissions close on 3 July 2025, with winners announced on 9 July.
Alexandra Ebert, Chief AI and Data Democratization Officer at MOSTLY AI, said the prize represents “a call-to-action for anyone with an interest in data and AI”.
“Open data access is key to unlocking AI’s full potential – but achieving that will require wider adoption of synthetic data tools,” Ebert said. “This challenge is about showcasing the power of privacy-safe data generation and making it accessible to all.”
The competition follows MOSTLY AI’s release of the first open-source toolkit for synthetic data generation. While participants can use this toolkit, it is not a requirement.
The challenge comes amid growing demand for AI training data and tightening privacy regulations. With traditional data sharing increasingly constrained, synthetic data – which mimics real-world data while stripping away identifiable information – is seen as a breakthrough solution.
MOSTLY AI, which raised $25 million in Series B funding and works with clients including Citi, Telefónica and the U.S. Department of Homeland Security, says synthetic data can help businesses and researchers share and scale data securely across industries.
Full details of the challenge, including data samples, scoring metrics, and eligibility, are available at: mostlyaiprize.com.