IA et fake news

Overview

The speaker argues that fears about “Lia” (an AI system) creating a crisis of confidence—and making it impossible to distinguish truth from falsehood—are overstated. They say misinformation existed long before AI, citing examples such as Photoshop and historical government propaganda.

Historical example of manipulation

As evidence that manipulated narratives persist over time, they reference Stalin-era doctored imagery—including a photo that was circulated in one form, where a secret police chief later disappeared from the official version. The point: propaganda and edited realities continued for decades, regardless of AI.

Core takeaway: credibility over blanket verification

Their main takeaway is that the solution isn’t simply trying to verify every piece of information. Instead, they emphasize focusing on:

• Sources
• Credibility
• Who posted the content
• The person’s reputation and ethics
• Whether the audience can reasonably trust the author

They contend that while AI-generated fake news may become abundant, it does not automatically replace the trust an audience places in specific creators.

Practical advice for creators/marketers in 2026

The speaker shifts to practical guidance for 2026. Rather than rushing to use tools designed to generate high volumes of content (they mention installing “Clo” and posting many times daily on LinkedIn), they recommend a different priority: build long-term audience trust.

They suggest adopting a publishing charter, including:

• Fact-checking
• Avoiding false promises
• Becoming the reliable person audiences turn to when overwhelmed

They conclude that the coming months will contain lots of “bullshit,” and that building trust requires more effort than mass-producing AI content—but it will differentiate someone when human-written and AI-written texts become harder to distinguish.

Personal note

They add a personal detail: they write a Sunday newsletter by hand.

Presenter(s)

• The speaker (name not provided in the subtitles)