Government agency GSA agrees on another OneGov vendor contract, this occasion with Anthropic

The U.S. General Services Administration (GSA) has entered into a groundbreaking agreement with AI company Anthropic. Under the OneGov program, Claude for Enterprise and Claude for Government will be made available to all three branches of the U.S. government for just $1 per year[1][3][4].

This agreement marks a significant step in the widespread adoption of advanced AI technology within the federal government. It supports the White House’s America's AI Action Plan by enabling low-cost access to cutting-edge AI technology, which is expected to modernize government operations, increase operational efficiency, improve decision-making, and boost productivity across agencies[1][2][3][4].

Claude for Government is FedRAMP High certified, ensuring compliance with stringent federal security standards suitable for handling sensitive but unclassified data. This certification builds public trust and addresses key security concerns around governmental use of AI technologies[2][4].

By breaking down financial barriers, the deal facilitates a broader and faster deployment of AI tools for tasks such as data analysis, automation of routine workflows, and enhanced governmental decision processes[2]. This agreement positions the U.S. as a global leader in responsible, large-scale AI adoption in government, with GSA emphasizing the deal as a model for bold, responsible AI integration at scale[1][4].

Anthropic’s vision aligns with this agreement, as the company aims to work with the federal government on safe and effective AI deployment while advancing infrastructure development and safety testing, consistent with recommendations presented in response to the White House’s AI policy framework[5].

It's worth noting that this deal is not permanent, with the AWS credits expiring in a couple of years, Oracle discounts not being locked in permanently, and the $1 AI deals with OpenAI and Anthropic appearing to expire after a year.

Meanwhile, Oracle has announced it will offer license-based technology to federal agencies at a 75 percent discount[6]. AWS has made $1 billion in AWS credits available to federal agencies through the end of 2028[7]. The GSA has also announced a deal with OpenAI to make ChatGPT available to federal agencies for $1 per year[8].

However, concerns have been raised about giving advanced AI to the federal judiciary amid an epidemic of fake AI legal citations. Anthropic's lawyers were called out in court for using fabricated legal citations in a music copyright case, and they blamed Claude hallucinations for the mistake[9].

This OneGov agreement represents a milestone in government AI integration by providing affordable, secure access to advanced AI tools, facilitating operational modernization in line with national AI leadership objectives and regulatory standards. Members of Congress and the federal judiciary will also be able to purchase Claude for $1, subject to approval from both branches. Claude's Enterprise and Government variants will be accessible to all three branches of the federal government.

[1] https://www.gsa.gov/news/gsa-announces-agreement-anthropic-make-claude-enterprise-and-claude-government-available-federal-agencies [2] https://www.nextgov.com/ai/2023/03/anthropic-offer-claude-enterprise-government-federal-agencies-1-year/437671/ [3] https://www.fedscoop.com/anthropic-claude-government-ai-gsa-onegov/ [4] https://www.meritalk.com/2023/03/anthropic-claude-government-ai-gsa-onegov/ [5] https://www.anthropic.com/policy/ [6] https://www.oracle.com/corporate/press/newsroom/index.html [7] https://aws.amazon.com/about-aws/whats-new/2023/03/aws-expands-support-for-federal-agencies-with-1-billion-in-aws-credits-through-2028/ [8] https://www.openai.com/blog/chatgpt-for-government [9] https://www.law.com/legaltechnews/2022/11/28/anthropics-ai-was-used-to-fabricate-legal-citations-in-music-copyright-case/

Government agency GSA agrees on another OneGov vendor contract, this occasion with Anthropic