// HACKER NEWS — CYBERSECURITY
When Google Cuts Off Access: Poland and the World
In brief: On September 1, 2026, Google disabled the Gmail inbox I had used for fifteen years. The reason: “spam,” without identifying a single message. That day I sent six complaints about Anthropic, the maker of Claude, to 107 recipients. I appealed five times, complained to three public authorities, went to Google’s offices, and sent three registered letters. Four companies stand behind the account, yet Google Poland is not a party to any agreement I have. Google’s servers in Poland are purchased by a company with five employees and no website. The bank resolved my dispute with Anthropic in eighteen days; in two weeks, Google did not identify a single message.
This article grew out of one question people ask when, overnight, they lose an inbox they have used since 2011: did I encounter one automated system’s mistake, or a way of doing business that can be documented through examples from Poland and around the world, one after another, with dates? The answer took two weeks and is set out below.
I used only publicly available material that anyone can verify without asking anyone for a favour: judgments and court orders, official announcements, company-register extracts, financial statements, reports by ombudsmen and regulators, and press accounts of people who lost access to their own accounts. When a fact comes from an official document, judgment, or register, I state it directly. When I know it from an article, blog, or law firm’s account, I say so in the sentence. Not every case described here concerns a blocked account; some involve competition, data, or advertising. They are included because they show how Google deals with courts and regulators, which matters to anyone trying to get an answer from this company and hearing only an automated system in return.
One more point about method. Several companies sharing an address or board members is a matter of public record. I draw no conclusions from that fact about anyone’s motives or the responsibility of any particular person. Readers are fully entitled to ask themselves why everything was arranged in precisely this way.
I am not a free user. I pay Google for business services, Workspace, and Cloud, and a Warsaw company within the group invoices me [18]. On September 1, 2026, between 15:49 and 15:53, I sent six messages from my private Gmail account to a total of 107 different addresses: 99 media addresses, four public authorities, two Anthropic addresses, and two private addresses. They were complaints about Anthropic, the maker of the Claude assistant and Claude Code tool, which had disabled services I paid for and answered my complaints with a bot. I sent them from Gmail because that was the address I used to log in to Claude, and Anthropic was copied on every message. That is why I did not use email on my own domain: otherwise their bots, which since April have passed my complaints back and forth dozens of times, would not know which account to associate the correspondence with. I did this deliberately because I know how these systems work.
Google’s automated system classified those six emails as spam and disabled the inbox the same day. The list of messages and recipients is attached to the pre-litigation demand and sits in the evidence file on this website [evidence]. Five appeals produced five automated refusals, repeated word for word, without a single concrete detail and without a human signature. On September 9, I filed a formal request through the form Google provides under the EU Digital Services Act; the submission received a number, but no answer arrived.
I later sent the same messages—the ones for which Google’s automated system called me a spammer—again from my business inbox on my own domain. No filter stopped them. Public authorities, companies, and individual newsrooms, including some outside Poland, are replying. People are replying, not a machine. The same text is correspondence to one email provider and a breach of rules to another, punishable by