The exciting and evolving world of Artificial Intelligence (AI) has taken an alarming turn, with Indiana University Bloomington researchers uncovering a botnet promotional racket exploiting the AI language model ChatGPT. Built by OpenAI, the platform has fallen into nefarious hands, promising golden cryptocurrency tips on X, courtesy of Fox8, a cluster of roughly 1,140 accounts.
This crafty botnet cleverly employed ChatGPT to spawn and engage posts, attempting to lure unknowing users to crypto-boosting websites. By churning out near-human messages that outline the charm of cryptocurrencies, Fox8 has shown how easily AI can be repurposed for scams.
However, the investigators luckily identified a quirky phrase “As an AI language model…” that ChatGPT often uses in response to specific commands, enabling them to manually examine suspicious accounts.
The revelation just marks the beginning of a potentially mammoth problem, according to Micah Musser, recognised for his expertise of AI-powered misinformation. Musser suggests that uncovered scams are possibly the tip of a detrimental iceberg, hinting towards myriad, more complex campaigns in the shadows.
Despite the clear guidelines outlined by OpenAI that prohibit the abuse of its models for scams, the study showcases the perils of uncovering such deceitful botnets when appropriately arranged, becoming invisible, and distorting algorithms to more effectively propagate disinformation.
Filippo Menczer, a lead professor at the University, conceded that it was sheer negligence on the scammer’s part that led to the discovery of the AI exploit. A shrewd adversary would make no such blunder, essentially becoming invisible to the typical scrutinization tools.
Spamming bots have long been a nagging issue in the cryptocurrency community, mostly targeting popular platforms like YouTube and X. Usually identifiable, they have managed to dupe victims out of millions, impersonating celebrities, and hosting deceitful giveaways.
Musk‘s promise to overpower these spam bots, especially after taking over X, appears to have fallen short, with bots reportedly running rampant. Menczer’s team no longer communicates its findings to X, citing a lack of responsiveness.
With Musk recently confirming the removal of X’s blocking feature, creators voice their concerns over a reduced capacity to filter out tricksters and impersonators. While AI usage in crypto promotion illustrates an intriguing development, it identifies the potential for manipulation and highlights the urgent need for robust detection and prevention mechanisms.
Source: Cryptonews
