Researchers: an ex-Florida deputy sheriff who received asylum in Russia has built a network of 160+ fake news sites with the help of ChatGPT and other AI tools
In 2016, Russia used an army of trolls to interfere in the U.S. presidential election. This year, an American given asylum …
Context & Ripple Effects
Related coverage traces a durable Russian-linked influence pattern: from organized troll operations in 2015 to botnets, site networks and paid trolls used around the 2016 election, and a reported revival of troll and bot activity ahead of the 2022 U.S. midterms. The reported site network suggests generative AI is being applied to the content-production layer of that familiar playbook.
The case also gained significance in subsequent coverage when documents reportedly showed direct ties between the former deputy and Russia's government. That connection makes the network relevant not merely as a misinformation operation but as potential influence infrastructure.
First-order effects
- A network of more than 160 purported news sites can supply a high volume of article-like material, making it harder for readers and intermediaries to distinguish independent reporting from coordinated output.
- ChatGPT and other AI tools reduce the labor required to draft and vary content, while putting greater scrutiny on the distribution channels, domains and accounts that carry it.
Second-order effects
- Search, social and advertising systems face a more domain-centric moderation problem: removing individual posts may not address a network that can continually publish new material.
- The operational bottleneck shifts from writing content to building audience reach and credibility, increasing the value of attribution research that connects sites, operators and distribution behavior.
Third-order effects
- If this model spreads, influence operations may rely less on conspicuous troll farms and more on persistent networks of seemingly independent publishing properties augmented by generative tools.
- The longer-term contest is likely to center on provenance and coordinated-behavior detection rather than text quality alone, because AI-generated material can be varied without changing the underlying operator network.
The trend: Generative AI is lowering the cost of operating influence-content networks while shifting detection toward infrastructure, provenance and coordination signals.