The Biggest Questions About AI
The map · 5 Humanity · 5.1 Truth, information, and democracy · 5.1.4

Information intermediation

Will a few assistants increasingly determine what people read, believe, notice, and consider politically possible?

Farrell and Gopnik's frame: LLMs are cultural technologies, like print or bureaucracy — they reorganize what a society knows and how. A few assistants intermediating everyone's information is a governance question wearing a UX costume.

View on the map → · Open in Browse →

What changed
October 2025–August 2026 · swept August 3, 2026 · editorial review pending
01

The shift from search to assistants got measured from both ends. On the demand side, chatbot news use is climbing (Reuters Institute: 10% weekly, 16% of under-35s) even as trust in the answers sits near 20%, and roughly half of US adults now use chatbots. On the supply side, publisher search traffic fell about a third in a year. The quality of what assistants relay is the open worry: the EBU/BBC study found 45% of AI news answers had significant issues, and owner-aligned reference tools (Grokipedia) diverge from Wikipedia precisely on politically sensitive topics.

Recent thinking
5 featured from 8 tracked · October 2025–August 2026 · all 8 chronologically →
EBU & BBC · European Broadcasting Union · 22 Oct 2025 report
News Integrity in AI Assistants
45% of all AI answers had at least one significant issue.

The 18-country, 14-language public-broadcaster study that is now the reference measurement of assistant news accuracy: nearly half of chatbot news answers had significant issues and 31% had serious sourcing problems, consistent across languages.

Reuters Institute for the Study of Journalism · Reuters Institute, Oxford · 16 Jun 2026 report
Digital News Report 2026
Weekly use of AI chatbots such as ChatGPT, Perplexity, and Google Gemini for news increased year-on-year from 7% to 10%.

The authoritative annual measurement of the search-to-chatbot shift in news: chatbot news use up to 10% weekly (16% of under-35s) while trust in chatbot answers sits at just 20% globally.

Saeedeh Mohammadi & Taha Yasseri · PNAS · 15 May 2026 paper
Selective divergence between Grokipedia and Wikipedia articles
Grokipedia shows a consistent shift toward referencing more right-leaning news sources compared to Wikipedia.

First peer-reviewed measurement of an AI-generated encyclopedia diverging from Wikipedia selectively on politically sensitive topics — concrete evidence that owner-aligned assistants can quietly reshape reference knowledge.

Gottfried, Anderson et al. · Pew Research Center · 17 Jun 2026 report
Americans and AI 2026: Chatbots, Smart Devices and Views on Impact
About half of U.S. adults now report using AI chatbots, up substantially from the summer of 2024.

Pew's baseline survey of US assistant adoption: roughly half of adults use chatbots, 42% to search for information and 13% for news — the demand-side numbers for how far intermediation has shifted.

Charlotte Tobitt · Press Gazette · 12 Jan 2026 news
Global publisher Google traffic dropped by a third in 2025
Google search traffic to publishers declined globally by a third in the year to November

Chartbeat data quantifying the supply side of the shift: a third of publisher search traffic gone in a year as AI summaries absorb clicks, hollowing out the economics of the content assistants summarize.

Additional relevant discussion (3)
Censorship in Chinese AI models can be undone, new research shows — Reed Albergotti · Semafor · 29 Jul 2026
Measuring political bias in Claude — Anthropic · 13 Nov 2025
Foundational reading (2)Large AI models are cultural and social technologiesFarrell, Gopnik, Shalizi & Evans, Science · 2025The Ethics of Advanced AI AssistantsGabriel et al., Google DeepMind · 2024
Previous5.1.3 Political power