The Honesty Tax: When Google's AI Decides What Matters and Washington Makes Honesty the Law
What happens when Google's AI decides what matters in your inbox, and Washington state law says your subject lines can't mislead? Cold email just got a lot more honest. Two seemingly unrelated developments — a Gmail algorithm change and a state supreme court ruling — now converge to create a double bind that makes most traditional cold email subject line tactics legally precarious and algorithmically irrelevant.
The tension is simple. Google's new "What Matters Most" inbox promises 3 billion users an AI that surfaces the emails they actually need. Washington's anti-spam law now says subject lines containing any false or misleading information — even about promotions, not just commercial nature — can trigger statutory damages. The email marketer caught between them faces a problem with no easy workaround: optimize for engagement and risk liability, or optimize for honesty and risk invisibility.
What "What Matters Most" Actually Means for Senders
Google's May 2026 announcement that it is rolling out "new capabilities" for an AI inbox fundamentally changes what email deliverability means. The company's stated goal is to let users "see what matters most" — but the mechanism matters more than the slogan. Google is now using Gemini to generate "contextual drafts" of replies before users even read the original email. It surfaces relevant Google Doc, Sheet, or Slide links "right next to your to-do." This is not a spam filter update. This is a fundamental redefinition of inbox hierarchy.
For cold email senders, the implication is brutal: your email now competes not just against other senders, but against an AI that decides it knows what the recipient cares about. If Gemini determines a draft reply to your email is unnecessary or irrelevant, your message may never reach the cognitive portion of the inbox. It might be summarized, deprioritized, or literally never read because the AI already handled it.
The privacy implications, as security analysts have noted, are substantial. But for email marketers, the more immediate concern is behavioral. When an AI reads your email and decides its relevance before the human does, the subject line becomes the only piece of content that matters for deliverability — and it must be both honest enough to survive legal scrutiny and compelling enough to survive algorithmic triage.
The Washington Trap: Subject Line Liabilities You Didn't Know You Had
Meanwhile, in April 2025, the Washington Supreme Court ruled decisively in Brown v. Old Navy that the state's anti-spam law covers all false or misleading information in commercial email subject lines — not just information that misleads about the commercial nature of the message. The case itself is instructive: Old Navy sent emails announcing a "50 percent off promotion was ending," but continued offering the same promotion days later. Plaintiffs sued, and the court held that the subject line itself violated the statute.
The court's reasoning is direct: "No person may initiate the transmission of a commercial electronic mail message to a Washington resident that contains false or misleading information in the subject line." The justices explicitly rejected Old Navy's argument that the law was meant only to prevent spam deception about commercial purpose. Instead, they held that the legislature deliberately targeted subject lines because they are "the two pieces of information consumers first glean when faced with the choice of deleting a message or engaging with its content."
Here is where it gets concrete for practitioners. The court said that receiving the email itself is the injury. Statutory damages do not require showing financial harm — only that a commercial email with a false or misleading subject line was sent to a Washington resident. And the law covers more than obvious fraud. Old Navy's defense that "banal hyperbole" like "Best Deals of the Year" could create liability was rejected by the majority, though the court did note that puffery doctrine — statements that are obviously exaggerated and not factual claims — might still provide some protection.
The Double Bind in Practice
Take a typical cold email campaign scenario. You are selling B2B software. Your subject line reads: "Quick question about your workflow automation." The recipient in Washington didn't ask for a question. The email is actually a pitch. Under the Washington ruling, that subject line could be construed as misleading because it implies a genuine inquiry that does not exist. Your "quick question" is actually a sales outreach. The recipient can now sue for statutory damages, and the court's reasoning suggests they might win.
But if you change the subject line to "Sales pitch: workflow automation software," your open rates plummet. And under Google's new AI triage, that subject line screams "low relevance" to Gemini, which will deprioritize it in favor of messages the AI judges to be things the recipient actually cares about.
The result is a squeeze: the subject lines that work for cold email are increasingly illegal, and the subject lines that are legal increasingly don't work.
What You Can Do: The Honesty Optimization Framework
This is not a problem you can buy your way out of. No deliverability tool can fix a subject line that a state supreme court has deemed illegal. No AI optimization can save a message that Gemini has flagged as low priority. But there are concrete steps that shift the balance back in your favor.
- Audit every subject line for factual claims, not just puffery. The Washington ruling explicitly preserves puffery as a defense, but only for statements that are obviously non-factual. "Best software ever" is puffery. "We can reduce your costs by 30%" is a factual claim. If you cannot substantiate it, do not put it in the subject line. This applies to urgency claims ("Last chance," "Ending soon") and personalization claims ("I saw your post about X" when you did not).
- Segment Washington recipients separately. The law only covers commercial emails sent to Washington residents. If you can identify recipients by state, you can run different subject line tests for Washington-identified contacts. This is not perfect — VPNs, mobile devices, and forwarding make geolocation unreliable — but it reduces your surface area of exposure.
- Reverse-engineer relevance for Gemini. Google's AI is trying to determine "what matters most" to the user. That means your subject line must signal genuine relevance to that specific person, not generic urgency. Instead of "Quick question," try something that references a concrete, verifiable fact about the recipient's company or role. "Saw your recent post on supply chain delays" works if it is true. If it is not true, you are back in Washington territory.
- Test subject lines against the "hypothetical plaintiff" standard. Before sending, ask yourself: if a Washington resident received this email and a plaintiff's attorney asked them "Did this subject line contain any false or misleading information?" — could they answer yes? If the answer is maybe, do not send it. The standard is not intent; it is falsity in the subject line itself.
The Strategic Implication Is Uncomfortable
What this convergence really means is that the old cold email playbook — high-volume, high-urgency, borderline-deceptive subject lines — is becoming legally untenable for anyone sending to Washington, and algorithmically untenable for anyone relying on Gmail. The escape route some practitioners are considering — ignore Washington, focus on other states, rely on low enforcement rates — is short-sighted. The Washington Supreme Court's reasoning is strong, and other states often follow. Google's "What Matters Most" feature is rolling out globally and defaults to enabled.
The more honest your subject line, the less likely it is to trigger statutory damages. The more relevant it genuinely is to the recipient, the more likely Gemini is to surface it. Honesty and relevance are now the same thing. The problem is that most cold email campaigns are built on neither.
So here is the unresolved question: what happens when your entire sales funnel depends on subject lines that cannot be both fully honest and sufficiently compelling to generate opens? Is the response to be more creative in ways that signal genuine value without exaggeration, or do we acknowledge that cold email's foundational tactic — the subject line that creates false curiosity — is entering its terminal decline?
The answer may determine not just your deliverability, but whether cold email remains a viable channel at all.