What you learn: MIT Sloan faculty is applying AI research to tackle complex and critical challenges, from creating more equitable organ transplant policies to addressing the backlog of asylum systems.
Artificial intelligence is not just about writing emails. It is also a powerful tool to address some of the most urgent and complex issues of society.
At Computing Research Symposium's 2025 MIT Ethics, researchers across MIT, including some of the MIT Sloans, explored how AI can be used for public goods.
Below are four ways in which MIT experts are studying responsible applications of AI that can be useful for social welfare.
Creating a more equitable organ transplant policy
For patients waiting for an organ transplant, time is more than just precious. It's a matter of life and death. However, today's policy setting processes for organ allocation are slow and often silent across the region, making it difficult to adapt in real time or assess the long-term impact of change, said Professor MIT Sloan and Associate Dean.
To accelerate this process, Bertsimas and his team developed a simulation algorithm that could evaluate new organ distribution policies that are about 1,000 times faster than existing methods. Using this simulation, policymakers can better understand how proposed changes affect wait times and mortality rates, such as consolidating wait lists across the region. The team also aims to build a website to explore the effectiveness of various policies, increasing transparency and speed in historically slow processes.
Watch: A fair and efficient kidney transplant allocation analysis
Identify where new implantation techniques can make the biggest difference
The future of organ transplantation may include emerging technologies such as xenografting (using animal organs) and organ cryopreservation (freezing organs for later use), which can significantly expand the donor pool. But even if science is working, logistics remains important. It is important to know where and when these technologies will be deployed.
Associate Professor MIT Sloan AI is used to model the dynamics of organ transplant supply and demand to identify the geographic and demographic gaps in which these innovations are most useful. For example, if the region regularly has a high organ disposal rate due to timing discrepancies between donors and recipients, cryopreservation may be particularly valuable there.
Watch: Towards fair and efficient organ transplants with longer storage times
Maintaining diversity in algorithmic decision making
Algorithms should help you make faster and more objective decisions. But what happens when everyone's algorithms begin to make the same decision? According to Assistant Professor MIT Sloan Information sharing across AI systems can inadvertently reduce diversity and make the market as a whole and society more vulnerable.
Get a resume screening as an example. If multiple employers use similar AI models trained on the same data, they could reject the same candidate, even if those candidates stand out against the individual human recruiters. The same is true with rental set software. If all landlords use the same tool, city-wide rents could rise in lockstep. Raghavan's work explores how algorithm uniformity leads to conspiracy-like outcomes, and how policymakers and engineers can design for diversity, as well as accuracy in automated systems.
Watch: Information sharing, competition, and conspiracy through algorithms
Fixing a broken asylum scheduling system
The US asylum process has long been criticized for its inefficiency, inconsistencies and long delays. And while much of the public discussion focuses on policy, Assistant Professor MIT Sloan Are you asking another question: Can better scheduling make the system more humane?
Freund's work applies operational research and AI to the US asylum adjudication process. In this process, the number of applications is far greater than the number of decisions. Traditional queuing systems like “First In, First Out” don't always make sense, he said. Some people benefit from delays because they can work while waiting, while others face serious challenges from long-term limbo. Freund's work highlights how algorithm scheduling can reduce backlogs, allocate resources more efficiently, and minimize harm to vulnerable populations.
Watch: Frontiers of equity efficiency in humanitarian immigration
Dimitris Bad Simah He is a professor of operations research, associate dean of business analysis and vice principal of open learning. His research interests include optimization, machine learning, and application probability. Swati Gupta He is an associate professor of operational research and statistics. Her research focuses on deep theoretical challenges in optimization and AI. Manish Ragavan I am an assistant professor of information technology. His research studies the impact of computational tools on society. Daniel Freund I am an associate professor of operational management. His research applies optimization, stochastic modeling, and revenue management techniques to transport, online platforms, and humanitarian immigration issues.
