Crypto Games ExtensionCrypto Games AppTelegram NewsSubmit
Anthropic taps Accenture to run third-party AI safety evaluationsImage Source: Engadget

Anthropic taps Accenture to run third-party AI safety evaluations

George Tsagkarakis 2 min read
Contents 3 sections
We may include affiliate links in our content, meaning we could earn a commission—or receive blockchain-based assets—if you click a link and make a purchase or take a specific action. Additionally, we use generative AI to help draft and refine our posts for clarity and grammar. All content is fact-checked and reviewed by a human editor before publication.

Anthropic is about to let a consulting firm sit in on its model training. Accenture will place third-party evaluators inside the AI company, and the two are each putting up at least $1 billion over the next five years to fund the arrangement.

That’s a lot of money for a role nobody has defined yet. Anthropic said itself that the embedded evaluator idea is new, with no settled standards on scope, on what access those evaluators get or on how they should report concerns.

The deal is the first concrete move under a proposal Anthropic CEO Dario Amodei laid out last week for slowing the pace of AI development. His three-step plan asks every “frontier AI” company to give “ongoing, employee-like access” to embedded outside evaluators. Their job is to confirm safety practices and commitments are being followed, report incidents and assess how well models are aligned.

What Accenture’s people will do inside Anthropic

Anthropic’s description is unusually specific about access. Accenture’s evaluators will “watch models take shape in training, follow the decisions that govern how those models are built and deployed and speak directly to employees,” the company said.

Anthropic’s case for Accenture rests on the firm’s record helping businesses and governments deploy AI across several industries. The company also said Accenture knows how enterprises use AI in practice, which it argues brings a perspective on Anthropic’s own models that the lab doesn’t have by itself.

Accenture’s day job, though, is consulting, and its client list now includes the company whose models it’s supposed to grade. Anthropic didn’t address that tension. It did say Accenture will evaluate AI companies besides Anthropic, so this isn’t an exclusive relationship.

Accenture won’t be the only one

Anthropic said it will bring in other companies beyond Accenture to evaluate its frontier models and plans to name them in the coming weeks. Until then, the only hard numbers in the arrangement are the ones on the check: at least $1 billion from each side, five years and a rulebook that hasn’t been written.

Share this article
George Tsagkarakis

George Tsagkarakis, known as Staycalm4now is a professional author in the crypto gaming industry since early 2018. He has experienced all the growth of Blockchain Gaming and helped multiple projects achieve their goals and established a player base. He is the co-founder of egamers.io and now the Founder and owner of CryptoGames.gg He is also the COO of MyStage, an…

More from AI Companies & Funding

Leave a Reply