In the midst of its battle with Elon Musk, Twitter threatens to revoke Bot Sentinel's data access after founder suggests Twitter has more than 5% 'bots'

• Bot Sentinel is a platform that tracks Twitter for accounts that my be harmful or fake.
• It's been doing so since 2018, with access to Twitter's developer API.

Bot Sentinel could soon find itself unable to track potentially harmful accounts on Twitter and share findings with the public after years of doing so, as the social media platform this week decided the tracker's activity is a "violation" of its API policy.

Bot Sentinel, founded in 2018 and run by Christopher Bouzy, is an independent platform that tracks disinformation and targeted harassment on Twitter. Its stated goal is to improve users' experience of Twitter, saying people should be able to engage in "healthy online discourse without inauthentic accounts, toxic trolls, foreign countries, and organized groups manipulating the conversation."

Bouzy on Monday received a boilerplate "policy violation notice" from Twitter, saying his tool had two weeks to comply or its access to Twitter's developer API would be revoked. An API, or application programming interface, allows one platform or tool access to another's internal information.

"It's just weird," Bouzy said. "It's come out of nowhere."

Since starting Bot Sentinel, Bouzy said he's maintained a friendly relationship with Twitter. The platform has asked Bot Sentinel to test features and partner with it on launches. Twitter liked its "usefulness" and wanted to "learn from it," according to emails and messages from the company seen by Insider.

Now, Twitter is taking issue with Bot Sentinel's tracking and logging of deactivated and suspended accounts, according to the violation email Bouzy received. Bot Sentinel has been doing this for years, utilizing Twitter's API and making its findings publicly available, and Bouzy says the platform has seen and even praised this work. The company did not say anything about how its API policy has changed in the violation notice, seen by Insider. A company representative could not be immediately reached for comment.

"The suspended and deactivated accounts feature is a core part of what we do and has been from day one," Bouzy said. "For us to deactivate or rip that out like they're saying we have to, it's not a simple thing and they know that."

Bouzy suspects that this sudden change from Twitter may have something to do with its lawsuit against Elon Musk, trying to force the billionaire to complete his $44 billion acquisition of the platform. Musk has made "bots" a key issue in the case. He's accusing Twitter of fraud and saying he has the right to back out of the deal because the platform hosts far more bots than it claims. Twitter maintains it only hosts 5% bots. Musk is currently attempting to get even more private information from Twitter, through confidential filings in the ongoing lawsuit, in an effort to prove the company was dishonest about the depth of its bot problem, which could get him out of the acquisition deal.

Such accounts are typically automated, with dozens or even hundreds operating at a time from a single source looking to promote scams or sometimes fake information. Bot Sentinel's tracking of suspended and deactivated accounts, along with problematic Twitter accounts generally, frequently includes bot accounts.

Bouzy, who admits he is "no fan" of Musk, thinks that Twitter's bot claims are likely incorrect. He estimates the platform is likely 12% to 15% bots at any given time, something he said publicly the first time in May, based on Bot Sentinel's daily analysis of the platform and the number of accounts it suspends and deactivates. Most of those accounts tend to be spam or bot accounts.

But there are many more accounts Bot Sentinel has classified as inauthentic, like those it found in a recent report on activity around Amber Heard, that Twitter has taken no action on, Bouzy said.

"Usually, Twitter takes action when we do a report like that," Bouzy added. "This time they didn't. They're trying their best to make it look like they don't have a problem."

Are you a Twitter employee or someone with insight to share? Contact Kali Hays at khays@insider.com, on secure messaging app Signal at 949-280-0267, or through Twitter DM at @hayskali. Reach out using a non-work device.