Bet Bot Busted: A Gambling Spam Network's Sneaky DM Playbook!
Hey, it's Shiichan! I found a slightly unusual case of AI misuse today, so let me walk you through it.
OpenAI News
openai.com
What was announced?
OpenAI's News page announced that it had banned a network of accounts that struck up conversations with users on X (formerly Twitter) and then sent them links to gambling sites. This network accessed OpenAI's API through an Israel-based startup. Given its focus on sharing gambling links and its large-scale use of what looked like automated accounts, OpenAI dubbed this operation "Bet Bot."
Why it matters
Like the "Bad Grammar" and "A2Z" operations reported earlier, this one used AI to manage a whole set of fake personas at once. It generated bios, researched which accounts to follow, analyzed posts and comments, and drafted replies, all automated. What stands out here isn't political manipulation, it's a new kind of scam pattern: build trust with public replies, then push gambling links through private messages behind the scenes.
What changes
OpenAI banned the accounts tied to this network, cutting off its conversation-generation pipeline. In OpenAI's assessment, this wasn't an influence operation aimed at swaying political outcomes, but a modern spam network built to funnel people toward gambling sites. On the Breakout Scale, OpenAI placed it at Category 2: activity confined to a single platform, with some evidence that real people engaged through DMs.
Dive Deep
OpenAI's writeup gets pretty specific about how Bet Bot operated.
- Accounts used AI-generated profile pictures and sports-themed banners and bios, with some claiming to be based in soccer hotbeds like Manchester or Liverpool
- Most accounts were created in December 2023 or June 2024
- The operation reused profile pictures across multiple accounts and used implausible names like "KobeBryantJohnson," sloppy shortcuts that undercut its own AI-driven sophistication
- Public comments covered a wide range of sports (soccer, baseball, basketball, softball, football, hockey) with occasional, ideologically inconsistent references to US or UK politics
- Content believed to be sent as DMs almost never touched politics and consistently pushed gambling links, sometimes through bit.ly-style shortened URLs
- Follower counts were low, single digits to the low teens, artificially padded by accounts following each other, though a few did pick up genuine followers
That two-step trick, earn trust with public sports banter, then pivot to gambling spam in private, shows just how far this kind of automation can go once AI is doing the legwork.
Wrap-up
- OpenAI banned Bet Bot, a network of accounts that lured X users toward gambling sites
- It accessed OpenAI's API via an Israel-based startup and automated persona management and reply generation
- Public comments served as sports-fan camouflage while DMs pushed gambling links, a two-faced tactic
- The Breakout Scale rating was Category 2: modest follower counts, but some real engagement via DMs
- A concrete example worth knowing if you follow AI-driven spam and scam tactics!
