scripod.com
When AI Decides You're a Threat — Brad Carson

Overview

Shownote

Highlights

Transcript

Chapters

Pins

When AI Decides You're a Threat — Brad Carson

Machine Learning Street Talk (MLST)

May 31
Machine Learning Street Talk (MLST)

Machine Learning Street Talk (MLST)

May 31
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Brad Carson was the Army's General Counsel, served two terms in Congress and was Acting Under Secretary of Defense for Personnel and Readiness. He now heads Americans for Responsible Innovation, the AI-policy advocacy group he co-founded. Keith Duggar spen...

Highlights

This podcast features a debate between Brad Carson, head of the AI-policy group Americans for Responsible Innovation, and Keith Duggar. Carson argues that AI development can be shaped and restrained, drawing on historical precedents like the Asilomar conference on recombinant DNA. Duggar challenges this view, focusing on the profound accountability problems created by opaque, probabilistic AI systems, particularly in military targeting.
00:00
Humans are essential for winning wars
04:52
Government oversight is more accountable than informal influence
09:40
Frontier AI companies have a public responsibility to be trustworthy.
12:24
Developers bear most, but not all, responsibility.
13:48
AI should be treated as a product, not a person
16:02
AI should be a product, not a person.
20:04
AI systems are unintelligible, making accountability impossible
28:24
AI gives a score without explaining why
28:48
Humans rubber-stamp machine decisions.
34:08
We can shape AI's future, just as we did with recombinant DNA.
42:39
Wars are ultimately won by people, not fancy equipment.
43:30
History shows human ground presence is essential.
56:31
Competition improves products and distributes power.
1:00:19
Unequal access to top models could create a dangerous bifurcation
1:20:12
The AI industry is its own worst enemy.

Chapters

From the Pentagon to AI governance
00:00
Regulatory capture vs Silicon Valley networks
04:52
Transparency and the Claude tier changes
07:56
Tort liability when AI tools cause harm
09:40
AI is a product, not a person
13:40
Children, suicide, and the suicide business
16:01
Opaque neural nets and the law of war
19:59
Probabilistic targeting and the death of accountability
25:54
The arms race fallacy: Asilomar and restraint
28:47
Talking to China: track 2 talks and chip leverage
34:02
Air power never wins: capital for labour
39:45
Anthropic vs the Department of War
43:29
Concentration, open source, and brain drain
51:29
DeepSeek, Chinese culture, and AI as diplomacy
1:00:18
Upskilling Congress and why public trust matters
1:12:25

Transcript

Brad Carson: So I remember when I was first elected to Congress 20 years ago. Now, the Congressional Management Foundation gives you this book of kind of like, how to be a congressman, how to run your office, what to expect day to day. And I remember readi...