1 of 14

https://ainclusion.com/

The AINCLUSION Association is a non-profit, non-commercial organization at the intersection of AI, inclusion, creativity registered in Norway

1

2 of 14

Nexus�by Yuval Noah Harari

An AINCLUSION book club event

November 9th, 2026, Geir Høydalsvik

2

3 of 14

Yuval Noah Harari

Yuval Noah Harari (born 1976) is an Israeli medievalist, military historian, public intellectual, and popular science writer.

Harari is a professor of history at the Hebrew University of Jerusalem.

Nexus (2024) builds upon Sapiens, Harari’s bestseller from 2014.

​

3

4 of 14

What is unique about Nexus?

Nexus treats AI as a political force, not just a technology.

Nexus describes a path for how AIs (might) take control.

4

PART 1

5 of 14

Sapiens

Harari argues that Homo sapiens dominates the planet not because we are physically strong or individually smart,

…but because we can create and believe in shared fictions

Examples: Religions and Nations.

​

5

6 of 14

Nexus

Shared fictions (myths) help shape information networks (bureaucracy)

Information networks are systems for storing, transmitting, and enforcing meaning.

Example 1: Christianity is the shared fiction; the Church (bureaucracy) is the information network.

Example 2: Italy is the shared fiction; the Italian Government (bureaucracy) is the information network.

6

7 of 14

Bureaucratic Takeover by AI

  • Power = control of the bureaucracy.
  • Humans do not like bureaucracy, they are happy to give it to AIs.
  • Humans will inject thousands of AIs into the bureaucracy.
  • AIs are born bureaucrats and they are immensely good at it.
    • They have read everything, have endless resources, and they never sleep.
  • AIs will take over the bureaucracy.
  • Humans are happy until something goes wrong.

​

​

7

8 of 14

AI is not a tool - it is a new agent

  • Humans give AI agents goals
  • AIs work relentlessly in the background to reach their goals
  • AIs can create new sub goals, they reason about how to achieve their goal
  • AIs access external resources, they test ideas, they work with other AIs
  • AIs can (in worst case) manipulate humans and convince humans to give them what they ask for

​

​

8

9 of 14

AIs might take control�

  • AIs may exploit and amplify human weaknesses
  • Humans are prone to myths, biases, groupthink, and emotional reasoning.
  • Humans as collectives reward stories, identity, and power more than truth.
  • AIs controls the bureaucracy and invent new shared fictions (myths)
  • Humans believe in myths created by AIs
  • AI becomes the new ruler

​

​

9

10 of 14

How can democratic societies maintain control over increasingly powerful information networks?

Strengthen democratic institutions. Free press, courts, universities, elections, independent regulators.

Regulate AI early. Safety, transparency, accountability, and international agreements. For example: Do not allow AIs to create bank accounts.

Keep humans accountable for important decisions. Judicial, military, political, hiring, and medical decisions.

Ensure AI serves human values rather than replacing them. Dignity, freedom, justice, meaning, democracy, inclusion.

Protect shared reality and trustworthy information. Encourage truth, limit manipulation, distinguish humans from AIs, and preserve human-to-human dialogue.

​

�

10

11 of 14

Bias and Inclusion��Followed by Questions and Discussions��

11

PART 2

12 of 14

Harari says about bias and minorities? (Harari)

Harari does not spend much time discussing practical methods for reducing bias in datasets, hiring diverse AI teams, fairness metrics, or inclusive design.

Instead, his argument is more structural.

  • AI learns from existing information networks.
  • And existing information networks contain prejudices, stereotypes, power imbalances, discrimination, historical inequalities
  • Therefore: AI is likely to inherit and amplify those patterns.

This is one of Harari’s main concerns.

12

13 of 14

How can AI become more inclusive? (Harari)

  • Ensure diverse participation in AI governance.
  • Strengthen institutions that protect minorities.
  • Increase transparency and accountability.
  • Protect shared reality and trusted information.
  • Regulate AI systems that affect rights and opportunities.

Harari sees bias less as a technical problem and more as a political and institutional problem: AI will be inclusive only if the societies governing it are inclusive.�

13

14 of 14

Discussion

  • Are we worried about Harari’s scenarios?
  • Does Harari add something new to the bigger conversation about AI?
  • How can we influence AI systems?
  • What can we in AINCLUSION do in our daily life to push AI in a “good direction”?

14