It's not exactly a game to train AI to spot legal issues.
Learned Hands is a game, and it does learn from users how to spot legal issues. The fun comes from winning points by giving correct answers in different factual scenarios.
But it's kind of like getting an honorary degree. The pay-off for players is bragging rights -- and maybe an inspirational message.
There is a higher reward for playing Learned Hands. David Colarusso, director of Suffolk University Law School's Legal Innovation and Technology Lab, says the game is designed to help consumers get answers to legal questions.
"It's an opportunity for attorneys to take their downtime to train machine learning algorithms to help access-to-justice issues," he told the ABA Journal.
Online legal services do not always match accurate answers to legal questions. For example, if a person searches for a "car accident" involving a "wife" and "kids," the search program may point to family law -- which would not help.
Colarusso said the Learned Hand project will build a dataset to create "algorithmically driven issue spotting." Players win points for discerning correct legal issues, and that teaches the computer.
Learned Hands is not the first computer game for lawyers, and it is not the first to help people learn about the law.
However, it is the newest to apply artificial intelligence and the "wisdom of the crowd" to help consumers find the right answers. Colarusso said it should spot issues correctly, at a minimum, 95 percent of the time.
Once the results are properly organized, the project team will release the data publicly for other researchers to use and train machines. The project is part of a collaboration with the Stanford Legal Design Lab and funded by PEW Charitable Trusts.