Mass AI Hub awards nearly $3.3 million to five applied AI projects
A Harvard robotics team plans to release its models openly. Other recipients will tackle coastal flooding, disease surveillance, worker skills and counter-drone coordination.
Listen to this story
The audio brief
Story brief
3 key pointsThe 2026 Applied AI Models Innovation Challenge is directing nearly $3.3 million across five Massachusetts projects, from industrial robotics to public-health surveillance and counter-drone coordination. The awards support applied research, not a set of newly launched products: Harvard’s Du says his team plans to open-source its video-generation models and inference code, while the other teams are developing forecasting, skills-assessment and sensor-fusion systems. For builders and operators, the portfolio signals state-backed investment in practical AI; delivery, adoption and the planned open-source release remain to be seen.
- 01
Harvard assistant professor Du’s team is developing video-generation models to help industrial robots learn tasks from demonstrations, with small and midsize businesses among the intended users.
- 02
The Atlantic Resiliency Innovation Institute is combining weather models, sensors and infrastructure-impact modeling to forecast local flood risks for energy infrastructure.
- 03
Boston University’s PandemIQ Llama supports disease surveillance and reporting through an open-access platform that the article says public health agencies worldwide already use.
Mass AI Hub has awarded nearly $3.3 million to five applied AI projects, with recipients celebrated at Harvard’s Kempner Institute. The 2026 Applied AI Models Innovation Challenge backs work ranging from industrial robots that learn from video demonstrations to tools for coastal flood prediction, disease surveillance, worker skills assessment and counter-drone operations.
Massachusetts Secretary of Economic Development Eric Paley joined representatives of the Massachusetts Technology Collaborative and Mass AI Hub at the recent ceremony, alongside Harvard Senior Vice Provost for Research John Shaw. Harvard described the gathering in an October 27 account. The university is a founding academic partner in the state-led Massachusetts AI Hub.
Teaching industrial robots through video
One award supports a team led by Harvard computer science assistant professor and Kempner investigator Du. The researchers will develop large video-generation models for industrial robotics. Their work focuses on creating detailed visual descriptions that could help robots learn industrial tasks from video demonstrations and adapt without manual programming.
The intended users include small and midsize businesses in manufacturing, logistics and recycling. The project’s goal is to make automation more flexible and affordable for those businesses. That is the ambition behind the funded research, rather than an announcement that a finished robotics system is now available.
Du said the team ultimately plans to release the models as open source, along with inference code—the code used to run them. The planned release is intended to let others connect the models to robotics platforms of their choosing, extending access beyond the research team.
We’ll ultimately plan to open source and release this model as well as provide some inference code that allows anyone to hook up these models to any type of robotics platform that they would like
Du, in his award video, as quoted by Harvard
Four other applications share the awards
- Atlantic Resiliency Innovation Institute is developing a coastal resilience platform to predict flood risks at a highly local level for energy infrastructure. The project combines enhanced weather models, sensor networks and modeling of infrastructure impacts, bringing those elements together in an AI-driven forecasting platform.
- Boston University’s PandemIQ Llama, part of its BEACON Platform, uses a specialized large language model for real-time global disease surveillance, risk assessment and automated epidemiological reports. The project aims to improve early outbreak detection and speed public health responses through an open-access system already used by public health agencies worldwide.
- Bristol Community College plans an AI-enabled system that assesses evidence of workers’ skills against industry frameworks. Its stated goals include stronger links between education providers and employers, support for hiring and workforce readiness, and opportunities for workers to build new skills.
- UMass Lowell Applied Research Corporation will use AI to coordinate multiple sensor systems for counter-drone operations. Its Autonomous Robotic Planning and Optimization for Intelligence Sensor Fusion platform aims to create a vendor-neutral interoperability layer, allowing robotics and defense systems to work together across suppliers.
Paley framed the awards as investments in responsible AI that addresses industry challenges and strengthens strategic sectors across Massachusetts. Shaw emphasized the collaboration needed to sustain that work, pointing beyond Harvard and Kempner to partner universities, affiliated hospitals and the private sector. The AI Hub itself is a broader initiative launched by the Healey-Driscoll administration to build the state’s position in applied AI.
Sources
- news.harvard.eduKempner hosts Mass AI Hub grant award celebration
Reader comments
Newest comments first. Replies stay oldest first.