Blog
You’re forty minutes into a vendor demo. The dashboards are colorful, the screens are slick, and somewhere around minute thirty, you hear the line every AI hiring vendor says now: “our platform is built on ethical AI.” Maybe it’s “bias-free.” Maybe it’s “fair by design.” The salesperson says it with the kind of confidence that’s supposed to settle the question.
You nod. You write it down. You move on to the next slide — because what else are you supposed to do? Ask for the audit? You’re not entirely sure what that would even look like, or what you’d do with it if they handed you one.
Here’s the uncomfortable part: a lot of the time, neither does the vendor.
“Ethical AI” Has Become Marketing Wallpaper
Every AI hiring vendor on the market today claims some version of fairness, ethics, or bias-mitigation. It’s in the homepage hero text, the sales deck, the one-pager your procurement team filed away.
It has become so universal that it has stopped functioning as information. When everyone claims it, the claim itself tells you nothing — it’s wallpaper, not evidence.
This isn’t a story about bad actors. Most vendors making these claims believe them. The problem is structural: “ethical” is a values word, and values words are nearly impossible to verify from the outside.
You can’t hold a vendor accountable to a standard you can’t inspect. What you can hold them accountable to is whether their claims are transparent — documented, audited, and published somewhere you can actually go look.
That’s a different question than “is this ethical?” It’s a better one, because it has an answer you can check.
When a vendor tells you their AI is fair, what would you actually ask to see?
The Date That Stopped Being a Deadline
For the last several months, anyone tracking AI hiring compliance has had a date circled: August 2, 2026, when the EU AI Act’s high-risk obligations — the ones that directly cover recruitment and employment AI — were set to become enforceable. It showed up in webinars, law firm alerts, and more than a few vendor sales pitches as the reason to act now.
In May 2026, EU institutions reached a political agreement to push that deadline back — to December 2, 2027, over a year later (source). The regulator missed its own deadline.
That’s worth sitting with, because it cuts against the way most compliance content gets written — including, frankly, the way we originally scoped this piece.
The instinct is to find the scariest date on the calendar and build urgency around it. But if even the EU AI Act’s own enforcement timeline can move by more than a year with a few months’ notice, a compliance date was never going to be a reliable reason to act. It was a deadline you could wait out.
Verification isn’t. You don’t need a law to force a vendor to show you their audit. You need to ask, and you need an answer that doesn’t depend on which government finished its paperwork on time.
That’s not true of every relevant rule, to be clear. NYC Local Law 144 has been enforced since July 2023, with civil penalties that accrue per violation, per day, for any employer using an automated employment decision tool without a current bias audit, public disclosure, and candidate notice.
That one hasn’t moved, and it isn’t going to. The lesson isn’t “regulation doesn’t matter.” It’s that some compliance obligations are durable and some are political footballs, and you can’t always tell which is which from the outside — so the safest assumption is to verify regardless of what the calendar says.
If the law that’s supposed to force transparency keeps moving, what's actually making your vendor transparent right now?
What Verification Actually Looks Like
Most TA leaders have never had to evaluate a vendor’s compliance posture before, because until recently, no one was asking.
That’s changing, but it doesn’t mean you need a law degree. It means you need a short list of specific things to ask for — not a feeling, an artifact.
Five questions do most of the work:
Who conducted the bias audit, and are they independent of the vendor? A vendor cannot audit its own tool and call it third-party.
Is the audit result public, or just available on request? “We can send you a summary” is not the same as a published dashboard.
Does the audit break results out by subgroup — race, sex, and intersectional categories — or just report an aggregate “no bias detected” claim?
Is there a human oversight mechanism built into the workflow, or does the tool make a final, unreviewed call?
Are candidates notified that an automated tool is being used, with a real opt-out — not buried in a terms-of-service page no one reads?
If a vendor hesitates on any of these, or answers in marketing language instead of specifics, you have your answer. Transparency either exists or it doesn’t. There’s no partial credit.
What We Do, Because We Were Asked to Prove It
We built this scorecard because we went through the exercise of answering these five questions about ourselves first, and we’d rather show you the answers than tell you to trust us.
CLARA is audited by an independent third party on an ongoing basis, for both bias and broader compliance, and the results are published on a public compliance dashboard — not sent on request, not summarized in a sales call.
We’ve been compliant with NYC Local Law 144 since before it was enforced, which means the audit, the public disclosure, and the candidate notification workflow are not new additions bolted on for a deadline. They were built in from the start.
We’re also built to the EU AI Act’s high-risk standards already, regardless of when the enforcement date formally lands.
That’s a deliberate choice: we’re a U.S.-based company, but a meaningful share of our partners hire across borders, and “we’ll get compliant when the deadline forces us to” isn’t a posture that protects anyone.
Being ahead of the requirement — not just current with it — is what lets us tell partners they’re covered no matter where they hire, not just where the law happens to currently reach.
None of this makes CLARA the only option. It makes CLARA one data point in a verification exercise you should be running on every vendor you evaluate — including us.
If you asked your current AI hiring vendor for their public audit dashboard tomorrow, what would they send you?
You Don’t Need a Legal Team for This — You Need Five Questions
Most mid-market TA teams don’t have an in-house compliance or legal function dedicated to vendor evaluation.
If you’re reading this, there’s a decent chance you are the closest thing your organization has to one — not because you asked for that responsibility, but because someone has to ask the questions and it landed on you.
You don’t have a legal team for this, and you don’t need one — but we’ve got you covered anyway.
The five questions above are things you can ask and evaluate yourself, in a single vendor call, with no specialized training. Where it gets harder is what happens after you have the answers: most of the time, you’re not the final decision-maker.
You’re the person who has to walk into a conversation with someone above you — a CEO, an owner, a board — and make the case for why this matters, or why a vendor failed the test.
That handoff moment is where most verification efforts quietly die. The checklist gets filled out, the insight stays in your head, and the leadership conversation never quite happens because there’s no clean way to bring it to them.
Verify It Yourself, Then Forward It Up
That’s why the Transparent AI Vendor Evaluation Scorecard isn’t just a checklist. It’s built to be filled out by one person in about twenty minutes — no legal review required — with a one-page summary at the front built specifically to be forwarded or printed for exactly that leadership conversation. Fill it out yourself. Then forward the summary.
It’s vendor-neutral by design. It doesn’t score CLARA against the criteria, and it isn’t built to make any single vendor look good — including us. Use it on the vendor you’re demoing next week. Use it on the one you already signed. Use it on us.
The next time a vendor tells you their AI is ethical, will you ask them to prove it — or write it down and move to the next slide?
Get the Transparent AI Vendor Evaluation Scorecard — a free, vendor-neutral tool built to help you verify what you’re being told, not just trust it.