Anthropic and Accenture Form Strategic Partnership for AI Safety
As part of a plan to slow down artificial intelligence development, Anthropic has selected Accenture as its first embedded evaluator.
Anthropic has announced the selection of Accenture as its first embedded evaluator, taking its first concrete step to implement CEO Dario Amodei's proposal to slow down the pace of artificial intelligence development.
Multi-Million Dollar Investment and Funding Plan
Both companies agreed to invest at least $1 billion over the next five years to build capacity in this area.
Due to the urgency of the process, Anthropic will directly fund Accenture's work, but argues that long-term funding should come from shared pools or government sources.
Embedded Working Model and Safety Tests
Under the partnership framework, employees within Accenture's artificial intelligence specialist unit, Faculty, will be embedded inside Anthropic.
This team will serve to test safety measures, subject models to red-team testing, and evaluate alignment with human values.
Expandable Collaboration and Other Discussions
This partnership is not exclusive, and Anthropic is continuing discussions with the non-profit research organization METR and other third parties as part of the process.