When Poulson resigned in August, he said he only planned to share his concerns about Dragonfly with those inside Google. But when Google didn’t respond to a group of human rights organizations that presented it with a letter arguing that Dragonfly is unethical and asking the company to kill the project, Poulson felt compelled to share his opinion with the public.

“I’m offended that no weight has been given to the human rights community having a consensus,” he said. “If you have coalition letter from 14 human rights organizations, and that can’t even make it into the discussions on the ethics behind a decision, I’d rather stand with the human rights organizations in this dispute.”

Regarding Poulson’s departure, a spokesperson for Google said, “It is our policy to not comment on individual employees.”

The revelation of Dragonfly provoked an immediate backlash within the company’s rank and file, who have high expectations for transparency from executives because of Google’s stated corporate values. One employee who’d been asked to work on the project decided to quit, another transferred teams, and internal forums were flooded with thousands of posts, comments, and emails debating the ethics of the project.

The controversy over Dragonfly followed closely on the heels of another ethical conflict at Google. In March, employees learned Google was working with the Pentagon to build artificially intelligent technology to be used in drone warfare. As with Dragonfly, some employees were shocked and angered by the news of the initiative, called Project Maven, and thousands signed a petition asking Google to cancel the contract. After a dozen engineers quit and cited Maven as the reason, Google agreed not to renew the contract when it expires this year.

After a two-week-long silence regarding Dragonfly, Google executives first addressed the company’s plans for China at an all-hands staff meeting held a month ago, which was interrupted when executives realized attendees were leaking details of the meeting to the press. At the meeting, Google CEO Sundar Pichai said the company is not planning to launch a search product in China in the near term.

Google executives and employees alike look down on leaks, and sources who spoke were concerned about the repercussions they might face for talking to the press. A source familiar with the situation told that Google has banned employees from livestreaming all-hands meetings on personal computers.

“We’ve been investing for many years to help Chinese users, from developing Android, through mobile apps such as Google Translate and Files Go, and our developer tools. But our work on search has been exploratory, and we are not close to launching a search product in China,” a spokesperson for Google said regarding this story.

But Poulson said even exploratory work done without the full knowledge of employees is a cause for concern. Following Project Maven, Google released a set of AI ethics principles that promised, among other things, not to “design or deploy AI” that “contravenes widely accepted principles of international law and human rights.”

“We can debate whether or not [Dragonfly] was going to be deployed, but that’s almost irrelevant because the AI ethics claim that they promise not even to design it,” Paulson said. “At what stage do engineers have a voice? … What I worry about is, once something is built and is ready to launch, the power has been transferred out of the engineers’ hands to a small group of people’s hands. You’ve effectively de-democratized the ethics of the development project.