Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Blog Pricing

Stephen Dubner

๐Ÿ‘ค Speaker
7195 total appearances

Appearances Over Time

Podcast Appearances

Freakonomics Radio
619. How to Poison the A.I. Machine

When Ben Zhao says that some artists reached out, that was how he started down his current path, defending visual artists. A Belgian artist named Kim Van Dun, who's known for her illustrations of fantasy creatures, sent Zhao an invitation to a town hall meeting about AI artwork.

Freakonomics Radio
619. How to Poison the A.I. Machine

When Ben Zhao says that some artists reached out, that was how he started down his current path, defending visual artists. A Belgian artist named Kim Van Dun, who's known for her illustrations of fantasy creatures, sent Zhao an invitation to a town hall meeting about AI artwork.

Freakonomics Radio
619. How to Poison the A.I. Machine

It was hosted by a Los Angeles organization called Concept Art Association, and it featured representatives from the U.S. Copyright Office. What was the purpose of this meeting? Artists have been noticing that when people search for their work online, the results were often AI knockoffs of their work. It went even further than that.

Freakonomics Radio
619. How to Poison the A.I. Machine

It was hosted by a Los Angeles organization called Concept Art Association, and it featured representatives from the U.S. Copyright Office. What was the purpose of this meeting? Artists have been noticing that when people search for their work online, the results were often AI knockoffs of their work. It went even further than that.

Freakonomics Radio
619. How to Poison the A.I. Machine

It was hosted by a Los Angeles organization called Concept Art Association, and it featured representatives from the U.S. Copyright Office. What was the purpose of this meeting? Artists have been noticing that when people search for their work online, the results were often AI knockoffs of their work. It went even further than that.

Freakonomics Radio
619. How to Poison the A.I. Machine

Their original images had been scraped from the Internet and used to train the AI models that can generate an image from a text prompt. You've probably heard of these text to image models, maybe even use some of them. There is Dolly from OpenAI, Imagine from Google, Image Playground from Apple, Stable Diffusion from Stability AI, and Midjourney from the San Francisco Research Lab of the same name.

Freakonomics Radio
619. How to Poison the A.I. Machine

Their original images had been scraped from the Internet and used to train the AI models that can generate an image from a text prompt. You've probably heard of these text to image models, maybe even use some of them. There is Dolly from OpenAI, Imagine from Google, Image Playground from Apple, Stable Diffusion from Stability AI, and Midjourney from the San Francisco Research Lab of the same name.

Freakonomics Radio
619. How to Poison the A.I. Machine

Their original images had been scraped from the Internet and used to train the AI models that can generate an image from a text prompt. You've probably heard of these text to image models, maybe even use some of them. There is Dolly from OpenAI, Imagine from Google, Image Playground from Apple, Stable Diffusion from Stability AI, and Midjourney from the San Francisco Research Lab of the same name.

Freakonomics Radio
619. How to Poison the A.I. Machine

In the case of an artist like Van Duyn, this might include her online portfolio, which is something you want to be easily seen by the people you want to see it, but you don't want sucked up by an AI.

Freakonomics Radio
619. How to Poison the A.I. Machine

In the case of an artist like Van Duyn, this might include her online portfolio, which is something you want to be easily seen by the people you want to see it, but you don't want sucked up by an AI.

Freakonomics Radio
619. How to Poison the A.I. Machine

In the case of an artist like Van Duyn, this might include her online portfolio, which is something you want to be easily seen by the people you want to see it, but you don't want sucked up by an AI.

Freakonomics Radio
619. How to Poison the A.I. Machine

So, Ben, I know that some companies, including OpenAI, have announced programs to let content creators opt out of AI training.

Freakonomics Radio
619. How to Poison the A.I. Machine

So, Ben, I know that some companies, including OpenAI, have announced programs to let content creators opt out of AI training.

Freakonomics Radio
619. How to Poison the A.I. Machine

So, Ben, I know that some companies, including OpenAI, have announced programs to let content creators opt out of AI training.

Freakonomics Radio
619. How to Poison the A.I. Machine

So Ben Zhao wanted to find a way to help artists fight back against their work being either forged or stolen by these mimicry machines.

Freakonomics Radio
619. How to Poison the A.I. Machine

So Ben Zhao wanted to find a way to help artists fight back against their work being either forged or stolen by these mimicry machines.

Freakonomics Radio
619. How to Poison the A.I. Machine

So Ben Zhao wanted to find a way to help artists fight back against their work being either forged or stolen by these mimicry machines.

Freakonomics Radio
619. How to Poison the A.I. Machine

In addition to identity theft, there can be the theft of a job, a livelihood.

Freakonomics Radio
619. How to Poison the A.I. Machine

In addition to identity theft, there can be the theft of a job, a livelihood.

Freakonomics Radio
619. How to Poison the A.I. Machine

In addition to identity theft, there can be the theft of a job, a livelihood.