Most people talk about AI at one of two volumes: panic or wonder. Skynet and mass unemployment on one side, a machine that will write your novel and cure your disease on the other. Neither side has stopped to ask what AI actually is.

Can AI Ever Be Human? Buy the book now: https://www.cuapress.org/9780813240862/can-ai-ever-be-human/

This lecture answers that question directly. AI is data, algorithms, and output: pattern recognition running on oceans of text and image and audio, nothing more mystical than that. I walk through the real difference between narrow AI (every system running today) and general AI (a machine that reasons across domains the way a person does, which doesn't exist yet), and I bring in Apple's 2025 paper "The Illusion of Thinking" to show what happens to frontier reasoning models once a problem gets hard enough: accuracy collapses, and the models use fewer reasoning tokens, not more. They give up.

From there the lecture turns to what AI can offer when it's rightly ordered: cancer detection rivaling radiologists, tools that restore speech to people with motor impairments, wildfire prediction models, and less fertilizer waste on farms. It also names the risks already here: biased facial recognition, hollowed-out work, surveillance, and deepfakes eroding trust faster than anyone can correct it.

The second half makes the harder argument. No version of this technology will ever become aware of itself or choose between good and evil, because a switch is not a conscience: Apple's 2026 discovery of "refusal neurons," single switches that collapse a model's entire safety alignment when suppressed, makes that concrete. The same limit holds for love. Love requires freedom and vulnerability a tool cannot possess, and no simulation of affection changes that.

This is Lecture 1 in a series drawn from my forthcoming book, Vigilant Hope: Preserving Human Dignity in an Age of Artificial Intelligence (Ascension Press). Lecture 2, "Can AI Ever Be Human?", picks up where this one ends.

Chapters
0:00 Two volumes: panic and wonder
0:35 What AI actually is: data, algorithms, output
3:16 Narrow AI vs. general AI
4:44 The Illusion of Thinking and the Tower of Hanoi collapse
6:26 AI in your day, without you noticing it
7:53 The promise: medicine, accessibility, environmental stewardship
10:04 The risks: bias, work, privacy, misinformation
13:23 What AI will never be: consciousness
14:57 What AI will never be: moral agency
18:29 What AI will never be: love
20:27 How to respond: understanding, evaluation, engagement
22:26 In practice: design, education, policy
24:04 A tool, not a god or a ghost

Subscribe for Lecture 2, where we press into what it actually means to be a person, and why that's something a machine cannot become.

How do we recover meaning in a fragmented world?
Short reflections on consciousness, conversion, and responsibility.

Read the full essays at Mediated by Meaning: https://mediatedbymeaning.substack.com

#meaning #philosophy #faith #psychology #jordanpeterson #catholic #lonergan #selfunderstanding