Advancing Artificial Intelligence and Machine Learning in the U.S. Government Through Improved Public Competitions

TL;DR

This paper discusses how the U.S. government can enhance AI and ML development through improved public competitions, addressing current barriers and proposing methods to increase participation, data security, and operational impact.

Contribution

It identifies key challenges in public AI competitions and offers strategies to improve their effectiveness for government-led AI advancement.

Findings

Public competitions face issues like poor data quality and limited engagement.

Strategies include data protection, easier participation, and model deployment.

Recommendations aim to boost AI innovation and stakeholder involvement.

Abstract

In the last two years, the U.S. government has emphasized the importance of accelerating artificial intelligence (AI) and machine learning (ML) within the government and across the nation. In particular, the National Artificial Intelligence Initiative Act of 2020, which became law on January 1, 2021, provides for a coordinated program across the entire federal government to accelerate AI research and application. The U.S. government can benefit from public artificial intelligence and machine learning challenges through the development of novel algorithms and participation in experiential training. Although the public, private, and non-profit sectors have a history of leveraging crowdsourcing initiatives to generate novel solutions to difficult problems and engage stakeholders, interest in public competitions has waned in recent years as a result of at least three major factors: (1) a lack of high-quality, high-impact data; (2) a narrow engagement focus on specialized groups; and (3) insufficient operationalization of challenge results. Herein we identify common issues and recommend approaches to increase the effectiveness of challenges. To address these barriers, enabling the use of public competitions for accelerating AI and ML practice, the U.S. government must leverage methods that protect sensitive data while enabling modelling, enable easier participation, empower deployment of validated models, and incentivize engagement from broad sections of the population.