DiscoverCold Takes AudioAI safety seems hard to measure
AI safety seems hard to measure

AI safety seems hard to measure

Update: 2022-12-08
Share

Description

Four analogies for why "We don't see any misbehavior by this AI" isn't enough.

https://www.cold-takes.com/ai-safety-seems-hard-to-measure/

Comments 
Top Podcasts
The Best New Comedy Podcast Right Now – June 2024The Best News Podcast Right Now – June 2024The Best New Business Podcast Right Now – June 2024The Best New Sports Podcast Right Now – June 2024The Best New True Crime Podcast Right Now – June 2024The Best New Joe Rogan Experience Podcast Right Now – June 20The Best New Dan Bongino Show Podcast Right Now – June 20The Best New Mark Levin Podcast – June 2024
In Channel
What AI companies can do today to help with the most important century

What AI companies can do today to help with the most important century

2023-02-2018:27

Jobs that can help with the most important century

Jobs that can help with the most important century

2023-02-1030:42

Spreading messages to help with the most important century

Spreading messages to help with the most important century

2023-01-2520:09

How we could stumble into AI catastrophe

How we could stumble into AI catastrophe

2023-01-1328:54

Transformative AI issues (not just misalignment): an overview

Transformative AI issues (not just misalignment): an overview

2023-01-0525:03

Racing Through a Minefield: the AI Deployment Problem

Racing Through a Minefield: the AI Deployment Problem

2022-12-2221:04

High-level hopes for AI aligment

High-level hopes for AI aligment

2022-12-1523:49

AI safety seems hard to measure

AI safety seems hard to measure

2022-12-0822:22

Why Would AI "Aim" To Defeat Humanity?

Why Would AI "Aim" To Defeat Humanity?

2022-11-2946:14

The Track Record of Futurists Seems ... Fine

The Track Record of Futurists Seems ... Fine

2022-06-3021:21

Nonprofit Boards are Weird

Nonprofit Boards are Weird

2022-06-2325:27

AI Could Defeat All Of Us Combined

AI Could Defeat All Of Us Combined

2022-06-0723:56

Useful Vices for Wicked Problems

Useful Vices for Wicked Problems

2022-04-1225:19

Ideal governance (for companies, countries and more)

Ideal governance (for companies, countries and more)

2022-04-0517:33

Debating myself on whether “extra lives lived” are as good as “deaths prevented”

Debating myself on whether “extra lives lived” are as good as “deaths prevented”

2022-03-2920:20

The Wicked Problem Experience

The Wicked Problem Experience

2022-03-0214:50

Learning By Writing

Learning By Writing

2022-02-2216:25

Defending One-Dimensional Ethics

Defending One-Dimensional Ethics

2022-02-1526:00

Future-Proof Ethics

Future-Proof Ethics

2022-02-0227:09

Stakeholder Management and Cost Disease

Stakeholder Management and Cost Disease

2022-01-2718:05

loading
Google Play
Download from Google Play
Castbox
Download from App Store
usUnited States
00:00
00:00
x

0.5x

0.8x

1.0x

1.25x

1.5x

2.0x

3.0x

Sleep Timer

Off

End of Episode

5 Minutes

10 Minutes

15 Minutes

30 Minutes

45 Minutes

60 Minutes

120 Minutes

AI safety seems hard to measure

AI safety seems hard to measure

coldtakes