Excerpted from a Bradley Arant Boult Cummings LLP Blog by Whitt Steineker and Slates Veazey
If an employee uses marijuana off the clock and the employer doesn’t test for marijuana, isn’t it much like the tree falling in the woods?
A recent Wall Street Journal report (“More U.S. Workers Are Testing Positive for Marijuana. Fewer Employers Are Concerned,” July 10, 2026) suggests that is the case, laying out data that most employment lawyers have felt anecdotally for years: Marijuana positivity rates are climbing, and employer appetite for testing is shrinking. Here’s what the numbers actually show, and what that means for employers who haven’t yet updated their drug-testing policies.
The Data
According to Quest Diagnostics’ 2025 workforce testing data — roughly eight million tests — the numbers break down as follows:
- Urine tests – marijuana positivity rose to 4.4%, up from 3.9% in 2021.
- Hair tests (which can detect use up to 90 days back) – about 15% were positive, a nearly 60% jump since 2021.
- Random hair screenings – 21% were positive, the highest of any category cited.
Why Employers Are Pulling Back
The Journal cites Fisher Phillips partner Todd Logsdon and the firm’s 2024 survey of roughly 1,000 employers; about half don’t test for cannabis pre-hire at all, largely because it shrinks the applicant pool. Of those that do test, 44% reported recruiting difficulties tied to it, and nearly a quarter considered loosening their policy.
Named examples of companies that have already dropped pre-hire marijuana testing, per the article, include Citigroup, AutoNation, Home Depot and Amazon.
The Legal Backdrop
Three things are converging here, and employers need to track all three separately because they don’t move in sync:
- State law patchwork. The article notes New York bars testing most job applicants and workers for marijuana, and that at least two dozen states have some form of employment protection for medical marijuana users. Mississippi, on the other hand, has employer protectionary language in its cannabis laws. There’s no single federal rule here.
- Federal reclassification. The article states the Trump administration began the process this spring of reclassifying marijuana as a less dangerous drug (moving it off Schedule I), which could eventually affect testing policy. But it is important to note that rescheduling of non-medical marijuana is a regulatory process, not a completed one to date, and the article itself only says it “could eventually affect” employer policy — it does not say how or when.
- Safety-sensitive carve-outs are holding firm. Trucking, construction and similar industries are described as unlikely to drop full-panel testing, and more likely to lean on hair testing specifically to catch longer-term “lifestyle” use rather than just recent impairment.
Best Practices
We’re regularly asked by employers what they can do to have an effective and compliant workplace drug policy, particularly with respect to the evolving legal status of marijuana. A few things to point out:
- Separate “safety-sensitive” from “everything else.” Bifurcating the policy by role rather than a blanket approach. DOT-regulated and safety-critical positions are a different legal and practical calculus than a marketing hire.
- Shift resources from pre-hire testing to on-the-job detection. Logsdon’s suggestion in the article — ditch the pre-hire cannabis screen and monitor for actual impairment instead — is worth taking seriously as a recruiting-competitiveness move, not just a legal one. To be clear, though, I’m not aware of a mature, legally battle-tested standard for “impairment testing” the way there is for threshold-based urine testing. That means an increased role for employers in educating themselves and their employees on signs of impairment and taking action when an employee shows one or more of those signs.
- Audit state law before finalizing any national policy. Given the patchwork described above, a one-size-fits-all national testing policy is a liability risk in states with applicant protections (New York being the clearest example named). This needs jurisdiction-specific review.
- Watch rescheduling, but don’t act on it yet. Track the DEA process, but don’t change policy based on an anticipated outcome.
- Don’t assume “less concerned” means “no policy.” The trend is toward normalization, not elimination of employer interest — increasingly, off-the-clock use is being tolerated in a way that on-the-clock impairment is not, mirroring how alcohol is generally treated.
For the full story, please click here.