Skip to content

Deconstructing Power in the Age of AI

Recently, we published a response to a UN positioning paper, have been chatting with folks about AI Literacy and are otherwise thinking about the affect AI might have on us, both positive and negative.

To be honest, I’ve had an intuitive opinion about the advancements of AI. I haven’t sat down and purposely contemplated. None of the thematics around this topic are new for people in edtech. When the printing press was invented in 1440, there was societal upheaval as people learned to read and took back power from their overlords and the Church.

We know that technology changes society and that we can’t always predict how. However, education is, in part, about helping people create, hold and disseminate power for themselves. Personal power helps people fulfil their psychological and social needs.

So in this post, I want to noodle out loud about power + AI.

French and Raven’s six forms of power: a brief overview

The world changes constantly and understanding power dynamics has become essential with the latest onslaught of technological advancements. Power, the ability to influence others’ behaviour and decisions, can take various forms. Developed by social psychologists John French and Bertram Raven in 1959, the six forms of power provide a taxonomy for categorising how power is used to influence. This video is a basic explainer of types of power, loosely based on French and Raven’s taxonomy:

  • Legitimate Power: Derived from a person’s role or position within an organisation or society. It rests on the idea that some individuals think they have the authority to make decisions and expect compliance due to their status. “State Action” is a form of legitimate power, so is “my boss said…”.
  • Reward Power: This form of power stems from the ability to offer rewards, be they tangible or intangible, to influence others’ behaviour. “Wealth” is reward power.
  • Coercive Power: In contrast to reward power, coercive power relies on the capacity to administer punishment or negative consequences to achieve compliance. “Physical force” is coercive.
  • Referent Power: Rooted in admiration, respect, and identification, referent power emerges when individuals are influenced by those they hold in high esteem. Influencers have “Referent Power”.
  • Expert Power: Expertise and knowledge form the basis of expert power. People are inclined to follow and accept guidance from those possessing specialised skills or information. Academia and science (used to) be examples of expert power.
  • Informational Power: Similar to expert power, informational power hinges on access to crucial data and information that others require. Possessing such information can lead to influence over decision-making. Visualising data is a way to use informational power.
Photo by Kalea Morgan on Unsplash

AI manipulating power structures

It stands to reason that machines can now be used to manipulate power structures. As AI systems become more integrated into various aspects of our lives, the dynamics of power are shifting. So how will AI fundamentally transform the way power operates?

  • Legitimate Power will be questionable: AI can now modulate our voices, create photo realistic imagery and spout text that sounds legitimate. We can no longer believe what we hear, see or read.
  • Reward Power will be narrowed: Anyone can offer rewards to influence behaviour, but as wealth inequality and AI-related IP grow, so too will the diversity of people and organisations able to use this power. A lot of AI is kept in black boxes with models and data that are owned and accessible by a small segment of the technology industry. If we don’t open up commonly used models, the reward power will start to narrow to the organisations and people who own them.
  • Coercive Power will be ever more dangerous: Biased AI and discriminatory AI will begin to suggest executing this kind of power over people based on faulty data and models. As “legitimate” organisations begin to use AI to decide who is or is not allowed inside, who does or does not have the appropriate credentials, who is or is not likely to commit crime, who is or is not suitable for employment, we will see bias and discrimination proliferate.
  • Referent Power will be wielded by machines: Not only will we have an even harder time identifying if a person on the internet is, indeed, an actual person, we will see machines subverting systems humans were using to establish trust over the network.
  • Expert Power will be redefined: AI systems possess the potential to acquire, process, and interpret vast amounts of data in a fraction of the time it takes for humans. As AI systems get better, people will begin to trust them more.
  • Informational Power will be amplified: We’ve sort of left the data visualisation phase of the Internet behind as it became boring and common place. Nowadays, we know that those who can scrape, analyse and interpret data are more empowered than those who can’t. Companies and governments that harness data-driven insights wield greater influence in shaping policies, marketing strategies, and public opinions. AI will help destabilise people’s ability to access and harness informational power.

A note about “Cognitive Liberty”

The concept of “cognitive liberty” refers to the right to control one’s own cognitive processes and inner thoughts. As AI and neurotechnologies develop, questions arise about the extent to which individuals can maintain this liberty when their thoughts and decisions are influenced by external algorithms and AI systems. Doug shared a great podcast episode about the risks AI brings to Cognitive Liberty, and you can listen to it here.

Educating for the Future

As we progress into the age of AI, we need to protect and enable people’s personal power. We need to ensure that AI-powered decisions are transparent, unbiased, and aligned with human values. We need to foster education and awareness about the implications of AI and its effects on power dynamics. We have to empower people to make informed choices.

French and Raven’s six forms of power are great for a foundational understanding of how influence operates. Our technological landscape has introduced new dimensions to these dynamics, and we need to help people learn to navigate changing power structures.

H/t

  1. Protecting Our Freedom of Thought with Nita Farahany. H/t Doug Belshaw
  2. The video recommended in Dense Discovery by Jenny Sahng

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.