Former Twitter security chief warns Musk is going fast and ‘breaking things’

Elon Musk’s rapid changes to Twitter put the safety of its most vulnerable users around the world at risk, including human rights activists, free speech advocates and marginalized people in countries. autocratic, according to the social network’s former trust and safety manager.

“People need to very carefully and carefully weigh the costs and benefits of using Twitter, given their personal security situation,” said Yoel Roth, who resigned as chief trust and safety officer. Twitter security. All things Considered host Ari Shapiro. “And that’s a terrifying prospect to accept, especially for many people who have spent the better part of a decade building a platform, an audience, and a community on Twitter.”

Since Musk completed his $44 billion purchase of Twitter in late October, the billionaire has turned the company and the platform upside down, wreaking havoc and confusion on the employees, users and advertisers he depends on.

Musk quickly transformed Twitter’s previous approach to what is and isn’t allowed. It reinstated accounts that had been banned for violating Twitter rules, including that of former President Donald Trump, which was banned after the US Capitol uprising on January 6, 2021.

Last week, Musk announced a “blanket amnesty” for many suspended accounts (although he also suspended Ye, the rapper formerly known as Kanye West, on Thursday after Ye posted an image of a cross swastika inside a Star of David). This week, Twitter quietly updated its online policies to say it no longer enforces policies against misleading COVID-19 claims.

Meanwhile, Musk has fired half of the company’s staff and issued an ultimatum asking the remaining employees to commit to a new “hardcore” Twitter or quit. This has dramatically reduced the number of people working to keep Twitter users safe. The cuts include contracted content moderators, the company’s human rights team and investigators working to tackle political manipulation and child sex abuse material.

“That’s the real risk. You can’t do this job with a skeleton crew,” Roth said.

Twitter, which no longer has a communications staff, did not respond to a request for comment. In a blog post signed this week by “the Twitter team”, the company said its policies have not changed and that it “continues its diligent work to protect the platform against hateful behavior, abusive behavior and any violation of the Twitter rules”. He said the trust and safety team “remains strong and well-resourced.”

Roth said Twitter has been rightly criticized for being too slow to change. But Musk’s rapid transformation of the small but highly influential social network is alarming.

“Instead of this perhaps too slow culture, Mr. Musk is introducing a culture of moving fast and, unfortunately, breaking as a result,” Roth said.

In Roth’s nearly eight years at Twitter, he’s seen the company go through a cascade of crises, from Russian interference in the 2016 election to the company’s unprecedented decision to ban Trump.

After Musk took control, Roth was one of the few top executives left at the company as the new owner fired senior management.

Musk on Twitter vs. Musk in private

Roth described a rift between Musk’s public persona as a brash and temperamental autocrat — Musk changed his biography to “Chief Twit” after he made the deal — and the seasoned executive with whom Roth interacted.

“Those cartoons weren’t true to my experience with him,” Roth said. “Most of the time during the weeks we worked together, when a situation arose and I explained the rules, I explained the factors influencing the situation and suggested a course of action aligned with our policies. . , he listened and, often, accepted this approach.”

For example, one of Musk’s top priorities was to bring back some controversial accounts, including Babylon Bee, a conservative satirical site that was suspended for misinterpreting a Biden administration official.

Roth and Musk discussed whether reinstating the account would result in a broader change to Twitter’s rules against gender errors or be a singular exception.

“[Musk] was ultimately convinced that taking this kind of one-time action would undermine Twitter’s rules and create gaps in enforcement consistency that would make Twitter a less trustworthy place,” Roth said. (Musk would go on to reinstate Babylon Bee’s account with others after Roth’s resignation).

But as the days passed, Roth found that wasn’t always the case. When Musk took over, Roth wrote “red lines” he didn’t want to cross, including breaking the law and publicly lying. An important element: it would only remain as long as decisions were made based on Twitter’s policies and principles.

“What matters to me, at the end of the day, is not the decision, but how the decision is made,” he said. “I would not like to participate in the weakening [Twitter’s approach to governing the platform] with capricious decision-making. And unfortunately, that’s what happened.”

Roth resigned from Twitter on Nov. 10, a day after Twitter’s botched rollout of an $8-a-month subscription plan giving users blue checkmarks. Verifications previously indicated that the company verified the identities of high-level users, but under the new program, Twitter made no effort to confirm that followers were who they claimed to be.

It sparked a flood of accounts impersonating politicians, celebrities and big brands, further inflaming relations with advertisers, who were already wary of the platform’s direction under Musk. Roth’s team had prepared a lengthy document warning how the feature could be used in this way and offering guardrails to mitigate those risks, but it was largely ignored.

Just the day before he left, Roth had appeared with Musk in a public audio chat on Twitter where they tried to reassure advertisers that the platform was still a safe place for their brands.

Advertisers flee, hate speech increases

Less than a month after Musk took office, half of Twitter’s top 100 advertisers appear to have stopped spending on the platform, according to data compiled by liberal nonprofit Media Matters for America.

Civil rights groups have documented an increase in hate speech on Twitter since Musk took over. The company said a trolling campaign launched on far-right message boards led to an increase in racist, homophobic and anti-Semitic slurs shortly after the deal was struck, but that it reinforced the law enforcement and banned many responsible accounts.

On Friday, Musk tweeted a graph showing that hate speech “impressions” — that is, how many times people have seen such tweets — are lower than the level before he took over.

Hate speech impressions (number of times the tweet was viewed) continue to drop, despite significant user growth!@TwitterSafety will publish data weekly.

Freedom of expression does not mean freedom of access. Negativity should and will have less reach than positivity. pic.twitter.com/36zl29rCSM

— Elon Musk (@elonmusk) December 2, 2022

But a new report from the non-profit Center for Countering Digital Hate found that the daily number of tweets containing hateful terms was “significantly higher” in late November, compared to the baseline before it bought Twitter. The number of tweets with the word n ​​tripled during this period, for example.

Musk disputed the CCHR report because “completely wrong” and said hate speech impressions are less than 0.1% of what is seen on Twitter. It has pledged to publish the hate speech impression rate weekly.

Roth said he and former colleagues were “heartbroken” to see what was happening to their efforts to keep Twitter users safe.

“What happens on Twitter can move markets, change elections and impact the safety of millions around the world,” he said. “More than anything else, people [who have left the company] worry about what will happen, given how important Twitter is to the world, if there isn’t a team left to do this kind of work.”