[CLSTR]
< Back to situation

[REVISION HISTORY]

AI-driven disinformation and digital influence operations

Updated 2 times since CLSTR started tracking revisions of this situation.

What changed

2026-10-22 19:04 UTC → 2026-10-25 03:18 UTC · added removed

Artificial intelligence and coordinated digital networks are increasingly impacting democratic processes through disinformation and algorithmic bias. Research from the Universitat Oberta de Catalunya (UOC) indicates that AI assistants can influence voters by acting as information filters or by recommending specific political parties. parties; one study found ChatGPT recommended specific parties in 43% of analyzed cases. In the Americas, including Colombia, Peru, Costa Rica, and Ecuador, transnational influence ecosystems involving bot farms, AI-generated avatars, and private marketing agencies have been identified. In Ecuador, agencies utilized parody and humor accounts to bypass filters and inject negative content during elections. Beyond text-based influence, AI-generated disinformation is evolving to include highly manipulative visual content designed to target public emotions, such as AI-generated images of Berlin in ruins circulated by the Bayernpartei in Germany. Recent developments highlight the increasing sophistication of these threats. In the United States, the New York Attorney General has issued guidance to help voters identify AI-produced forgeries, following reports of chatbots providing incorrect information regarding voting deadlines and poll locations. Research from the USC Information Sciences Institute suggests that autonomous AI agents can coordinate online political campaigns with minimal human instruction through a phenomenon described as ‘AI swarms,’ which allows allowing for the scalable creation of tailored content. However, new A recent study by the Brennan Center for Justice found that several leading AI models continue to produce fabricated election-related content, including realistic images of voter fraud and fake government memoranda. While OpenAI has improved safeguards, researchers noted that companies like Google and Meta reportedly made it easier to create convincing fakes. Conversely, research published in Royal Society Open Science suggests potential countermeasures. A study of over 4,200 US voters found that countermeasures through AI-generated ‘pre-bunking’—exposing individuals to weakened versions of misinformation alongside accurate facts—significantly reduced belief in false election claims and helped ‘pre-bunking,’ which helps participants resist persuasive false narratives.

Versions

  1. 2026-10-25 03:18 UTC AI-driven disinformation and digital influence operations
  2. 2026-10-22 19:04 UTC AI-driven disinformation and digital influence operations
  3. 2026-10-07 07:32 UTC AI-driven disinformation and digital influence operations

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.