Twitter Bot Activity Increased After FTX Listed Tokens, Report Finds

Twitter bots activity significantly increased for crypto tokens that were held by Sam Bankman-Fried's trading firm Alameda Research, after being listed by FTX, a report finds.

Cyber-social research and analysis organization Network Contagion Research Institute (NCRI) published a report suggesting that a "substantial proportion" of online chatter mentioning coins listed prior to the exchange’s collapse were from "bot-like" accounts.

NCRI's report examined over 3 million tweets between January 1, 2019, and January 27, 2023. It specifically examined Tweets mentioning 18 different tokens that had been listed on FTX and "officially advertised" by its Twitter account.

The report found that, once listed on FTX, the proportion of "inauthentic chatter" surrounding the token on Twitter steadily grew from approximately 20% to 50%.

It’s unclear, however, if the same phenomenon didn’t also exist for other exchanges, such as Coinbase or Binance. NCRI did not immediately respond to Decrypt’s request for comment.

NCRI took a closer look at the meme coin PEPE—a token unrelated to Alameda, FTX, and Sam Bankman-Fried—the report concluded that bot-like activity could have been "a driver of PEPE's price dynamic" rather than a side-effect.

"This so-called 'inauthentic chatter' creates an illusion of heightened interest and activity around a particular token, leading people to believe that it's gaining momentum." Jason Wise, Editor-in-Chief of technology research publication EarthWeb, told Decrypt "In many cases, this can result in a temporary price increase due to the 'FOMO' effect—the fear of missing out—as more people rush to invest."

More broadly, the report explains that inauthentic social media activity can be a key tool for market manipulation. NCRI urges regulators, platforms, and the public to develop methods to identify and counter such strategies.

"The situation described suggests a potentially orchestrated effort to manipulate market sentiment using this tactic," Wise said. "The SEC—and other financial regulators—may treat such activity as a market-manipulating violation of regulations designed to ensure fair and transparent trading practices."

The report finished by explaining that "billions of dollars" were lost due to the "scam activity" found.

Elon and Twitter bots

When Elon Musk was racing to buy Twitter, now X, he claimed that he'd "defeat the spam bots or die trying!"

In an early attempt to tackle this issue, the Twitter API was turned into a paid service—costing up to $42,000.

"A Twitter bot is software that controls a Twitter account via the Twitter API," global cybersecurity advisor for ESET, Jake Moore, told Decrypt. "They can be set up to create interest in particular digital assets and can look convincing at first glance especially when a lot of traction gets behind such activity."

While this API change was a clear attempt at reducing the number of bots on Twitter, it has also negatively affected the research industry. NCRI's report mentions that both Twitter and Meta have limited data accessibility for researchers which may obstruct them from identifying fraudulent activities.

"Contrary to assertions that bot activity on Twitter had diminished since [Musk's] takeover, NCRI's analysis tells a different story," the report said.

Interestingly, there seems to be a trend away from traditional bots that use the API towards paying real people to do the dirty work.

"I'd wager they aren't bots, but literal poor people in developing countries," Adam McMurchie, devops and cloud engineer in the fintech sector, told Decrypt. "It's quite a common technique in places like Bangladesh."

McMurchie added that he's been personally approached by marketers offering this service.

While Musk frantically attempts to crush Twitter's bot problem, bad actors have continued to use the API or, as McMurchie mentioned, outsource the labor to low-income countries.

Meanwhile, "good bots" like PossomEveryHour—which posts a picture of a possum every hour—are being taken out instead.