Accenture

Garrett Bernson, last seen on China Talk in his role as deputy CDAO at the State Department.

He went to the then Department of Defense for a little bit, now makes a triumphant return in his new role as chief AI officer at Accenture Federal Services.

We're going to talk what the national security community needs besides just API access to wonderful models in order to really change how it does its mission.

Garrett, welcome back.

Thank you.

Thanks for having me.

And thank you to Accenture for bringing us this episode.

All right.

Appreciate your time.

Now, coming up, Anthropic says Accenture will evaluate its Frontier AI models as it looks to ease concerns about AI's potential to cause catastrophic harms to humans.

We have more on that story next.

This is Bloomberg Tech.

Reports this weekend also said the company's planning to stage its IPO in November, even as it faces heightened public anxiety about AI's potential to cause catastrophic harms.

Anthropic CEO Dario Amodei has been one of the most prominent voices calling for a slowdown of AI's development at the frontier.

Meanwhile, Anthropic said Accenture will provide independent evaluation of its Frontier AI models, with each committing at least a billion dollars over five years to building capacity in this area.

Accenture shares actually shot up after market Friday when it broke.

Here to do that with us is Bloomberg News' US equities reporter.

He is Matthew Griffin joining us here in studio, and we're gonna start with Accenture.

Yes.

Well, Carol, Accenture is getting a boost from all of this talk about AI safety that we've seen over the last couple weeks, and specifically from this announcement late on Friday that they are going to embed Accenture evaluators in Anthropic to try to make sure those models are safe.

We don't know a ton of details yet.

We also-

Well, apparently they're going to be making sure the AI models are aligned with human goals.

So it does at least sound more interesting, uh, than Office Space if, you know, a little vague for now.

Disney is reportedly bucking the theme park industry slump with price cuts and promotions.

Accenture was boosted by its Anthropic partnership.

HP's preliminary FY27 planning assumes MSD industry PC volume decline.

Critical Metals finished up 38.6% after Trump's Friday Greenland security agreement announcement.

And one of the things that's been proposed by Dario Amodei and agreed to by others is this idea of putting an evaluator embedded inside the company to report out on the state of affairs when it comes to alignment and containment.

And so over the weekend, I think, or maybe Friday, Anthropix and Evaluator Embedding Strategy took on some form and they announced a partnership with Accenture.

which is a major player in enterprise consulting, even for AI.

They've built an enormous business providing AI consulting to enterprises.

It's not easy.

I did that at Accenture as well for the whole thought leadership group and defining what those pillars were for each person.

And of course, they all wanted to say innovation.

Well, here we go again.

Yeah.

But that's important because that's where it lives.

You know, going back to my Accenture days, I was in tech and, And there were a lot of big tech companies working with Accenture, like Google, SAS, Oracle, Workday, Adobe, et cetera.

But we were all in tech.

De vuelta con Amber.

Anthropic ha seleccionado a Accenture como su primer evaluador externo integrado en una nueva iniciativa de seguridad, un movimiento significativo hacia la verificación por terceros del comportamiento de los modelos, en lugar de la autocertificación.

Señala una tendencia más amplia en la industria.

A medida que aumenta la presión regulatoria, los laboratorios están recurriendo a actores independientes fuera de sus propias paredes.

That means the rest is falling into ARM's and AMD's lap and going, thank you.

Anthropic expects to invest at least $1 billion over five years with Accenture to evaluate frontier AI models.

OpenAI, meanwhile, is reportedly looking at enormous increases in revenue and computing spending in the coming years.

Money goes somewhere.