Employer brand is often discussed as a function of what we in talent acquisition can control: marketing copy, candidate experience, and career pages. But in practice, it is usually an expression of the broader corporate brand, and that brand is shaped at the top. Sometimes a single high-profile decision can do more to define a company than years of employer branding campaigns.

Friday’s clash between Anthropic and the Trump administration is a case in point.

A quick recap: The Pentagon and Anthropic were negotiating a defense contract worth up to $200 million to deploy the company’s AI technology in the U.S. Defense Department’s classified network. The deal very publicly fell apart over Anthropic's reticence to allow the use its products in domestic surveillance and autonomous weapons. The administration ordered agencies to stop working with Anthropic, and the Pentagon said it would designate the company a “supply-chain risk” after talks broke down over guardrails on how its AI could be used. This is potentially a huge blow to the company, because the move could threaten Anthropic’s access to the key infrastructure necessary to conduct its business.

Within hours, OpenAI swooped in and announced that it had signed the deal that Anthropic would not. They later amended that agreement to address one of the sticking points that originally cost Anthropic the deal: the technology cannot be used to mass-surveil American citizens, though that assurance is unlikely to satisfy everyone.

So who made the right call? We will not know for a while. But from an employer brand standpoint, both companies just sent very clear signals. Anthropic signaled that it is willing to absorb political and commercial pain rather than risk what it views as misuse of their products. OpenAI signaled that it is pragmatic and willing to do what it takes to close the deal. Everyone likes a winner.

Those are two very different messages to current and future employees. The reaction to OpenAI’s deal was swift: internal dissent and an open letter signed by hundreds of employees opposing the militarization of AI.

We saw the same dynamic last year in Big Law. The Trump administration targeted major firms with executive orders that were widely viewed as retaliation for representing clients and causes Trump opposed, while also attacking diversity policies and political associations the White House disliked. The orders threatened security clearances, access to federal officials and buildings, and other business with the government, putting firms under intense pressure to choose between resistance and accommodation. Some fought the administration in court, while others opted to settle through agreements that included major pro bono commitments.

Months later, the jury is still out on which approach was more effective from a business standpoint. A Bloomberg Law analysis found that the five targeted firms that did not cut deals saw average headcount decline 2.8% from March 1 to September 8, while the nine firms that settled saw average headcount decline 4.9%. It is a useful data point, but far from conclusive given how many factors shape headcount decisions.

Whether they did it for moral or pragmatic reasons, these decisions have changed the way that some potential associates and even partners view the firms as employers.

In a few months’ time, some of these decisions may matter less to the businesses themselves. Companies can adapt, replace lost opportunities, and repair strained relationships. But the impact on their employer brands will last.

— David

P.S. If you found this newsletter valuable, chances are your colleagues will too. Feel free to forward it along—and if it landed in your inbox by way of a friend, you can subscribe here to get the next one directly.

Sponsored Content

Reduce Time to Hire and Increase Signal at High Volume

High-volume hiring often leads to longer time to hire, overloaded engineering teams, and inconsistent screening results. But it doesn’t have to. Join CoderPad to learn how leading talent teams are reducing screening time by up to 90 percent, lowering cost per hire, and shortening time to hire by a week or more. Hear how one customer redesigned their technical screening process to increase early-stage signal and confidently scale engineering hiring without adding recruiter headcount.

Featured Story

The Donkey Is Passing: AI and Jobs

Raghav Singh (who will be delivering a session on The Hidden Cost of AI in Hiring at the ERE Recruiting Innovation Summit in May) uses a 1925 New York Times lament about donkeys being replaced by machines as a mirror for today’s AI panic, where “white-collar obsolescence” has become the default storyline. He separates task automation from full job elimination, pointing out that past waves of productivity tools shifted work and changed roles more than they erased employment. The bigger question is not whether AI replaces people, but whether institutions, companies, and policymakers can manage the transition so productivity gains translate into real wages and stability instead of a more fragile labor market. (ERE)

Conferences

ERE Recruiting Innovation Summit

Atlanta, GA
May 5-6, 2025

The ERE Recruiting Innovation Summit is a practitioner-led event built for talent acquisition leaders who want real answers, not hype. Now that the updated agenda is live, this is your chance to explore the sessions, speakers, and conversations shaping the future of recruiting.

Join us for practical sessions, meaningful peer connection, and live Ask Me Anything office hours with speakers. If you care about where recruiting is headed and want ideas you can put to work immediately, you belong in this room.

(ERE Recruiting Innovation Summit)

Webinars

Detecting and Defending Against Candidate Misrepresentation

March 26, 2026 | 2:00 PM EST | 1 Hour

In this webinar we will tackle the rise of fake applicants, from AI-generated resumes and deepfake interviews to third-party test takers. We’ll cover practical tactics such as screening “traps,” interviewer training, verification tools for technical and non-technical roles, and legally defensible documentation. (ERE)