We finally have an ‘official’ definition for open source AI

OSI releases Open Source AI Definition (OSAID) to standardize open source AI criteria.

: The Open Source Initiative (OSI) has released version 1.0 of the Open Source AI Definition (OSAID), establishing criteria for open source AI. This definition helps align policymakers and developers, detailing requirements for AI models such as data transparency and modification freedoms. Meta and other companies face criticism for labeling proprietary models as 'open source,' sparking debate about data access and licensing. OSI acknowledges the need for updates as AI and its legal context evolve.

The Open Source Initiative (OSI) has unveiled version 1.0 of its Open Source AI Definition (OSAID) to provide a standardized definition for open source AI. Designed after years of collaboration, the OSAID describes criteria AI models must meet, including transparency in training data and the capacity for modification, aiming to establish consensus among policymakers and developers.

Although the OSI lacks enforcement power, it intends to identify AI models falsely labeled as open source according to the new definition. Meta, a significant player in AI, faces scrutiny because its models don't fully comply with OSAID standards, sparking disputes over data transparency and licensing expectations, even though it participates in drafting discussions.

Critics argue that the OSAID falls short in addressing proprietary training data licensing and intellectual property concerns. The OSI plans to revise the definition as AI technology and related legal matters develop, having set up a committee to monitor its application and propose further amendments. This collective effort seeks to more accurately reflect the complexity of today's AI landscape.

open sourceAI Definitiontechnology