BRIEFING.log
ACCESS GRANTED
Post
CheatGpt: When AIs Are Easily Influenced—Three Thought-Provoking Expert Tests - Didier Mayeur
Journal Du Net
Conference
Give AI enough convincing sources, and suddenly ponies qualify for the Olympics. Welcome to my kind of stress test. In its article on AI manipulation, Le…
Give AI enough convincing sources, and suddenly ponies qualify for the Olympics. Welcome to my kind of stress test.
In its article on AI manipulation, Le Journal du Net highlights my work on black hat SEO and Generative Engine Optimization, drawing on our discussion at SEO Summit Paris.
At the centre of the story: my Aquaponey experiment. The article describes how published content helped a fictional Olympic narrative appear in AI answers, and how fellow SEO professionals ended up portrayed as Aquaponey specialists.
Alongside my experiments, the article examines a classroom test involving brand recommendations and Reboot Online’s campaign to influence AI rankings through expired domains. It also addresses methodological limits and the ethical questions these tests raise.
My reason for exposing these weaknesses is simple: a confident answer deserves scrutiny, especially when someone else controls its sources.
Read the full article on Le Journal du Net (in French) :
https://www.journaldunet.com/intelligence-artificielle/1545355-cheat-gpt-quand-les-ia-se-laissent-influencer-trois-tests-d-experts-qui-donnent-a-reflechir/
In its article on AI manipulation, Le Journal du Net highlights my work on black hat SEO and Generative Engine Optimization, drawing on our discussion at SEO Summit Paris.
At the centre of the story: my Aquaponey experiment. The article describes how published content helped a fictional Olympic narrative appear in AI answers, and how fellow SEO professionals ended up portrayed as Aquaponey specialists.
Alongside my experiments, the article examines a classroom test involving brand recommendations and Reboot Online’s campaign to influence AI rankings through expired domains. It also addresses methodological limits and the ethical questions these tests raise.
My reason for exposing these weaknesses is simple: a confident answer deserves scrutiny, especially when someone else controls its sources.
Read the full article on Le Journal du Net (in French) :
https://www.journaldunet.com/intelligence-artificielle/1545355-cheat-gpt-quand-les-ia-se-laissent-influencer-trois-tests-d-experts-qui-donnent-a-reflechir/
IMG_ARCHIVE
1 files
/root/classified/seo-conferences-media/cheatgpt-when-ais-are-easily-influenced-three-thought-provoking-expert-tests-288/
