Featured series 
Harvard Impact Labs awarded $1.4 million to launch four new research projects that will tackle urgent social problems. Led by faculty from Schools across Harvard, these projects involve collaborations with leaders in the public, private, and social sectors to design, test, and scale solutions to four pressing challenges:
Paying workers what they’ve earned: Each year, employers fail to pay U.S. workers billions of dollars due to wage theft. This new project — led by Daniel Schneider and David Weil — will address this problem by collaborating with labor departments in New Jersey, Maine, and New York City to develop AI- and data-driven models to help them enhance enforcement and better allocate their resources. By developing tools that predict where violations are likely to occur and testing interventions to deter violations, this lab will generate new approaches that will boost family economic security, level the playing field for honest businesses, and make government more efficient. These tools will ultimately be made available to any interested city or state.
Transforming math education in low-resource settings: Millions of children in the U.S. and around the world lack foundational math skills, and teachers are often stretched too thin to provide the individualized instruction that students need. In partnership with Microsoft and Eedi, this new project — led by Sharad Goel, Emma Brunskill, Joscha Legewie, Teddy Svoronos, and Ying Xu — will design and rigorously test low-cost generative AI-powered tools aimed at transforming math teaching and learning across the world. The tools will give teachers real-time insights into individual students’ challenges and offer tailored suggestions for teachers to use in the classroom, while also empowering students to engage in self-paced, interactive learning. The team will test these new tools in classrooms in the U.K. and in the Global South, and the outputs of this project will be made open-source and shared through education and AI research networks for educators both in the U.S. and globally. Through this work, the team hopes to make personalized, high-quality math learning accessible to every child, everywhere.
Supporting good decisions in the legal process: Every day, lawyers and judges rely on “risk scores” to help make decisions that change people’s lives and maintain public safety, but these scores only predict the cost of inaction rather than fully weighing the benefits and risks of different options. To address this gap, this new project — led by D. James Greiner and Kosuke Imai — will develop and test “triage scores,” a new, evidence-based and AI-driven decision tool. The project will work with Norfolk Commonwealth Attorney’s Office and Philadelphia Legal Assistance to compare decision making using these triage scores to both decision making with no scores and decision making with currently used “risk scores.” The goal is to improve public safety, justice, and family well-being by helping decision makers understand the likely impact of their actions. If successful, triage scores could replace widely used risk scores in many domains, leading to better law enforcement, better service decisions, and better downstream outcomes.
Making lasting improvements in education: Governments and philanthropies around the world invest billions of dollars in education innovation pilots to tackle challenges in teaching and learning, but even the most successful pilots struggle to scale and persist. This new project — led by Asim Ijaz Khwaja, Tahir Andrabi, and Jishnu Das in collaboration with the Centre for Economic Research in Pakistan — will identify the conditions that enable effective education innovations to endure, spread widely, and make a lasting positive impact on hundreds of thousands of students’ lives. The team will partner with Khyber Pakhtunkhwa Elementary and Secondary Education Department in Pakistan, which is responsible for the education of 6 million students, to co-design and evaluate low-cost interventions that significantly increase student test scores. The researchers will ultimately share what they learn about organic scaling to improve teaching and learning around the world.
