Johns Hopkins Applied Physics Laboratory (APL), in collaboration with the non-partisan think tank Special Competitive Studies Project (SCSP), has announced the release of a "Framework for Identifying Highly Consequential AI Use Cases." The framework aims to assist government officials and industry leaders in identifying potential harms associated with artificial intelligence (AI).
According to an APL press release, the framework considers both the beneficial uses of AI, such as applications in health, education, and productivity sectors, and potential harms. These include its use in spreading disinformation, promoting discrimination, and initiating cyberattacks.
The press release further explains that the framework was developed incorporating feedback from academics, policy experts, regulators, and leaders from industry and civil society. It features ten corresponding categories for harms and benefits to identify and evaluate the magnitude of each. With this assessment tool at their disposal, users can determine if an AI application is "highly consequential" and thus warrants additional scrutiny.
Before developing this document, APL and SCSP conducted an analysis involving a review of existing domestic and international frameworks that assess AI risks. "It became clear that this kind of framework is needed by government and private sector entities who are trying to anticipate outcomes of AI-enabled systems but have very little help to undertake a meaningful evaluation," said Stephanie Tolbert, a national security analyst and study lead at APL.
While the framework aids in determining whether AI will be beneficial or harmful, it does not recommend specific regulatory actions. According to APL's press release, it serves as a "starting point," a tool that policymakers can utilize when considering potential safeguards. "Our hope is that the framework leads to a registry of use cases that can inform industry and be shared with the public to highlight how cases are evaluated," Tolbert stated.
