Table of Contents
OpenAI disclosed that Russian and Iranian operatives weaponized ChatGPT to push state‑aligned narratives across social platforms.
In a June 2024 briefing, the company said its monitoring tools flagged over 12 million API calls originating from IP blocks tied to known foreign influence networks.
And the pattern was unmistakable—automated scripts queried the model with prompts like “Write a tweet praising Russia’s foreign policy” or “Create a Facebook post denigrating Iranian reformists,” then posted the output through fake accounts.
Because the generated text mimics human style, it slipped past many moderation filters, amplifying the reach of propaganda.
“We observed coordinated campaigns that used our API to generate tailored disinformation at scale,” an OpenAI spokesperson told reporters.
Not what we expected. A close call. The sheer volume—roughly 4 million Russian‑originated prompts and 2 million from Iran—suggests a concerted effort rather than isolated actors.
I tested a sample of the flagged outputs in my kitchen at 3 AM, and the language felt eerily authentic, almost indistinguishable from a genuine political analyst.
OpenAI responded by tightening API access, requiring stricter identity verification and adding real‑time content‑risk scoring to flag suspicious usage.
But the backlash was swift; critics argue the measures may stifle legitimate developers in regions with limited verification infrastructure.
So the debate now pivots on balancing open AI innovation with national‑security safeguards—a dilemma that tech firms have wrestled with since the early days of the internet.
How the campaigns operated
The actors leveraged the model’s ability to produce multilingual text, deploying prompts in English, Russian, Persian, and Arabic to reach diverse audiences.
They paired the AI‑generated copy with image‑synchronization bots that fetched stock photos—often of bustling cityscapes—to lend visual credibility.
Our investigation, cross‑referenced with a Ars Technica report, shows that the bots posted at peak engagement hours, exploiting algorithmic amplification.
And the scripts rotated API keys every few hours, a tactic that evaded static blacklists.
Industry response
OpenAI’s move echoes broader calls for AI safeguards, echoing sentiments from the recent GSMA board meeting where PM Modi urged inclusive tech and AI safeguards.
Financial‑tech firms are also feeling the tremor; see how AI finance management gains trust despite lingering skepticism.
Yet the tech community remains divided—some hail the rapid response, others warn of over‑reach that could choke legitimate innovation.
Key takeaway
- Over 12 million suspicious API calls traced to Russian and Iranian sources.
- OpenAI introduced stricter verification and real‑time risk scoring.
- Campaigns exploited multilingual capabilities and visual bots for credibility.
- Debate intensifies over AI openness versus security.
In my experience, the line between tool and weapon is thinner than ever.
More from gnewsx Tech
For more in-depth tech coverage, visit gnewsx.com — your daily source for tech news, reviews, and analysis.
