Alphabet’s chief executive that the spam-filtering program trained on billions of messages may be “hitting the wrong target,” following reports that email from Republican fundraisers was ending up in spam at higher rates than their Democratic counterparts.
In a letter to Sundar Pichai, the chair of the F.T.C., Andrew Ferguson, cited reporting and complaints from conservative consultants that messages with links to WinRed are being rejected by Gmail but not messages with links to ActBlue, adding that the automated delivery of Gmail messages was having partisan consequences.
The chair warned that filters that prevent Americans from seeing political appeals or making a donation can violate the FTC Act’s ban on unfair or deceptive practices and that the agency could investigate and act on such cases if necessary.
Google has explicitly told reporters that it relies on a mix of signals — user spam reports, sender reputation, volume patterns and technical authentication including SPF, DKIM and DMARC — to determine delivery. That’s alongside the same toolbox that Microsoft Outlook, Yahoo Mail and most major email service providers use.
But those machine‑learning filters are probabilistic and opaque. The swap false positives for false negatives; tweak a model in an attempt to reduce the junk and you might inadvertently catch the legitimate big mass mailers. Since companies won’t disclose their own threshold levels, the resulting problem is visible only from the effect, not the cause.
Political campaigns depend heavily on bulk email for both fundraising and mobilization. Vendor polices, list hygiene, Authorisation configuration and recipient interactions contribute to varying rates; a “same” message can be signed and sent from different organizations, however with different deliverability results.
Consulting firms like Targeted Victory say that blanket filtering constitutes political bias; tech platforms and independent observers cite operational reasons — bad list maintenance, recycled IP addresses, or campaigns routed through low-reputation vendors are common culprits.
Conservatives have had legal and regulatory complaints before: the Federal Election Commission had thrown out a previous complaint over Gmail, and federal courts tossed out own similar lawsuits leveled by party organizations. In a separate case, a judge rebuked an effort to enlist the F.T.C.’s investigatory powers as retaliation against a media watchdog.
Yet the political moment is distinct. The F.T.C. now has a chair chosen by the current administration, and Congress has demonstrated increased interest in platform transparency. Regulators in Europe are also seeking greater disclosure from big tech companies under regulations such as the Digital Markets Act.
As far as deliverability is concerned, the classic advice still applies: adhere to SPF/DKIM/DMARC, clean up your list of inactive addresses, don’t blast unexpectedly from unfamiliar IPs, and stick to reputable ESPs with strong sending reputations (think: Mailchimp or SendGrid).
For Google and other platforms, regulators will probably push for more clarity explaining how political mail is treated and for appeal mechanisms enabling senders to figure out why a message was filtered. We can expect transparency reports and third‑party audits to be on the horizon.
Email delivery matters for donations, volunteer recruitment and the rhythm of modern campaigns; Pew Research Center surveys show digital tools are playing key roles in political organizing and fund-raising. As a result, perceived or actual bias in those channels can have an outsize political impact.
How Alphabet responds — technical fixes, improved sender education or formal transparency measures — will influence the larger argument about platform power, free expression and regulatory oversight for months to come.