Researchers fear what a Musk acquisition might mean for Twitter research data

Much has been written about Elon Musk’s bid to acquire Twitter, an effort which, despite substantial backing from Morgan Stanley and the approval of Twitter’s board, stands on unsure footing at present.

Reporting and punditry have focused on the security implications of the proposed acquisition, as well as Musk’s potential approaches to content moderation and, on a related subject, his understanding of the concept of “free speech.” But another consequential aspect of the deal has received considerably less attention: how Twitter’s data access policy for research might change under a Musk regime.

Twitter hasn’t always had a cozy relationship with researchers. However, in recent years, the social network has made strides in providing access to its archives at a time when rivals have taken the opposite step. In January 2021, Twitter claimed that academic researchers were one of the largest groups using its API.

Some researchers are concerned that Musk doesn’t share the same commitment to open data access, particularly considering the vitriol he’s shown in the past toward reporting that paints his ventures (including Tesla) in an unflattering light.

In 2018, Musk pledged to — but didn’t ultimately — build a website to rate the “core truth” of articles and journalists in response to reports on crashes involving Tesla cars, Tesla labor issues and his relationship with Wall Street.

Mor Naaman, a professor of information science at Cornell Tech, envisions a future in which Musk becomes hostile toward researchers exposing Twitter’s “challenges and deficiencies.”

“I am pessimistic that Twitter will continue to strive for accountability as a privately held company under Musk,” Naaman, who’s worked with Twitter data since 2009, told TechCrunch via email. “I do not believe research like we have done on [former President Donald Trump’s] Stop the Steal campaign — and the data we collected from Twitter and made available to other researchers, used in 12 different papers since last year — would be allowed to happen under Musk. Second, I cannot imagine internal teams that scrutinize the ethics and bias of the company’s systems will continue to function well, let alone publish their findings publicly.

“If they do continue to publish, these publications will have a much harder time overcoming the already existing suspicion around the corporate-friendly bias nature of platforms putting out their own research papers.”

Among other promises, Musk has said that he plans to “defeat spam bots” on Twitter — seemingly alluding to the malicious accounts that parrot misinformation and perpetuate scams. But not all bots are harmful, Orestis Papakyriakopoulos, a a postodoctoral researcher at Princeton, pointed out to TechCrunch via email.

For example, he said, researchers have used automated accounts to study whether shifting people’s attention to accuracy can increase the quality of the news they share on Twitter.

“I do not know how the acquisition will affect researchers’ data access, since it is not clear what will change on the platform. Musk’s statements until now have not focused to data collection,” Papakyriakopoulos said. “Of course, removing automated accounts will reduce the ability of researchers to perform field experiments on the platform. But honestly, automated activity through APIs is important for commercial and news agencies, so I do not believe that this is an action that will be done in the end.”

Megan Allison Brown, a research scientist for the Center for Social Media and Politics at NYU, shares Naaman’s pessimism about Musk’s intentions where they concern academic studies reliant on Twitter data.

She, like Naaman, points to Musk’s track record of “blocking critics” and “reducing transparency” at other companies he’s led — Tesla being the prime example. The company, which dissolved its PR department in 2020, has never offered data to prove Musk’s claim that cars driven on Autopilot, Tesla’s proprietary driver-assistance system, are safer than those driven by human drivers. The Los Angeles Times detailed how the automaker responds to whistleblowers with firings, censorship, lawsuits and even the hiring of private detectives.

“While Musk has made a lot of claims about increased platform transparency, reduced content moderation, etc., it’s hard to know where Musk’s ideas and the practical realities of running the platform will meet. … In practice, academics could potentially be blocked from using the platform or the API if they publish research critical of the platform,” Brown said via email. “All in all, time will tell whether Musk’s Twitter acquisition is good or bad for researcher data access, and there are many reasons it could be both.”

Twitter’s ironic reward for having accessible data has been heightened scrutiny, noted Zeve Sanderson, the founding executive director at the Center for Social Media and Politics and one of Brown’s colleagues. Twitter leadership has remained supportive even in light of negative findings, but the change in ownership shows how tenuous the situation is — especially given reporting that Musk floated the idea of job cuts as he negotiated his Twitter bid. Musk could argue, facetiously or not, that maintaining an API that fuels negative press cycles isn’t in the best interest of shareholders.