Facebook’s reason for banning researchers is untenable


When Facebook says On Tuesday, it suspended the accounts of a group of New York University researchers, which made the company’s hands seem bound. The team has been crowdsourcing data on political ad targeting through browser extensions. Facebook They have been warned many times that it is not allowed.

“For months, we have been trying to work with New York University to provide their three researchers with the precise access rights they requested in a privacy-preserving manner,” wrote Facebook Product Management Director Mike Clark. In a blog post“We are taking these actions to prevent unauthorized crawling and to protect people’s privacy in accordance with our privacy plan ordered by the FTC.”

Clark refers to Consent order impose Depend on Federal Trade Commission In 2019, and fined $5 billion for privacy violations. You can understand the company’s dilemma. If researchers want one thing, but powerful federal regulators need other things, then the regulator will win.

It’s just that Facebook is not in that dilemma, because the consent order does not prohibit what researchers have been doing. Perhaps the company’s behavior was not to maintain the government’s favor, but because it did not want the public to know one of its most closely guarded secrets: who would see which advertisements and why.

FTC penalties stem from Cambridge Analysis scandal. In this case, nominal academic researchers can access Facebook user data and their friends’ data directly from Facebook. These data notoriously ended up in the hands of Cambridge Analytica, which used it to micro-position on behalf of Donald Trump’s 2016 campaign.

New York University Project Advertising observer, The way of working is very different. It cannot directly access Facebook data. Instead, it is a browser extension. When users download the extension, they agree to send the ads they see, including the information in “Why did I see this ad?” Widgets for researchers. The researchers then infer which political ads are aimed at which user groups-Facebook does not have public data.

Does this arrangement violate the consent order? As you can imagine, the two parts of the command can be applied. Section 2 requires Facebook to obtain user consent before sharing data with others. Since Ad Observer relies on users who agree to share data, not Facebook itself, it doesn’t matter.

Princeton University computer science and public affairs professor Jonathan Mayer (Jonathan Mayer) said that when Facebook shares data with the outside world, it “has a certain obligation to supervise this data sharing relationship.” “But there is no question about whether users want to leave and tell third parties the order of what they see on Facebook.”

Facebook spokesperson Joe Osborne (Joe Osborne) admitted that the consent order did not force Facebook to suspend the researcher’s account. Instead, he said, Article 7 of the decree requires Facebook to implement a “comprehensive privacy plan” to “protect the privacy, confidentiality, and integrity of user data.” It is Facebook’s privacy plan, not the consent order itself, that bans what the Ad Observer team has been doing.Specifically, Osborne said that researchers repeatedly violated parts of Facebook Terms of Service It states, “You may not use automated means (without our prior permission) to access or collect data in our products.” The blog post announcing the account ban mentioned 10 crawls.

Laura Edelson, a PhD student at New York University and co-founder of Ad Observer, rejected the suggestion that the tool is an automated tool. scraper fundamental.

“Fetching is when I write a program to automatically scroll through a website and let the computer drive the way the browser works and what is downloaded,” she said. “This is not how our extension works. Our extension rides with users, and we only collect data on the ads shown to users.”

Bennett Cyphers, a technical expert at the Electronic Frontier Foundation, agrees. “There is no really good, consistent definition of scraping,” he said, but when users choose to record and share their personal experiences on the platform, the term is strange. “It doesn’t seem like Facebook can control it. Unless they say that users are violating the terms of service by recording their interactions with Facebook in any way.”

In the end, whether the extension is really “automated” is a bit digress, because Facebook can change its policies at any time—or, based on existing policies, it can simply give researchers permission. So the more important question is whether advertising observers really infringe on anyone’s privacy. Facebook spokesperson Osborne said that when the extension delivers ads, it may expose information about other users who do not agree to share its data. For example, if I install an extension, it may share the identities of friends who like or comment on ads.


Source link