A look at OpenAI's “red team” of 50 academics and experts, hired in 2022 to look for issues such as toxicity, prejudice, and biases in GPT-4 before its release
Microsoft-backed company asked an eclectic mix of people to ‘adversarially test’ GPT-4, its powerful new language model
Financial Times Madhumita Murgia
Related Coverage
- Interview: Why AI Needs to Be Calibrated for Bias Undark Magazine · Dan Falk
- A professor hired by OpenAI to test GPT-4 says there's ‘significant risk’ of people using it to do ‘dangerous chemistry’ Insider · Sawdah Bhaimiya
- Sam Altman says OpenAI won't train GPT-5 for “some time” and the letter to pause AI development missed “most technical nuance about where we need the pause” The Verge · James Vincent
- I spoke with Madhumita Murgia for her article in the Financial Times about my experience red-teaming OpenAI's GPT-4. … Roya Pakzad
- I am so excited to share that I was recently interviewed by Madhumita Murgia, a reporter from Financial Times, about my involvement in redteaming OpenAI's latest and most advanced AI system, GPT-4. … Boru Gollo
- OpenAI confirms that it is not currently training GPT-5 and doesn't have plans to do so in the near future. — This is in line with my current belief that GPT-4 is pretty close to where LLMs will peak. … @Riedl@sigmoid.social · Mark Riedl
Discussion
-
@csetgeorgetown
@csetgeorgetown
on x
“The reason why you do operational testing is because things behave differently once they're actually in use in the real environment.” CSET's Heather Frase spoke to @FT's @madhumita29 about red-teaming GPT-4. https://www.ft.com/...
-
@hatr
Hakan
on x
“After Andrew White was granted access to GPT-4, the new artificial intelligence system that powers the popular ChatGPT chatbot, he used it to suggest an entirely new nerve agent.” https://www.ft.com/...
-
@madhumita29
Madhumita Murgia
on x
Soon after GPT-4 came out I spoke to @paul_rottger for a story, who sparked the idea to speak to as many “red-teamers” as possible. Here's the fruit of that reporting which hopefully gives some insight into the inner workings of this complex technology - and how it could go wrong…
-
@betweenmyths
Ken Craggs
on x
@emollick “White [@andrewwhite01] told the Financial Times he had used GPT-4 to suggest a compound that could act as a chemical weapon...The chatbot then even found a place to make it.” https://www.ft.com/... @FT #AI #GenerativeAI #GPT4 #OpenAI #OpenAIChatGPT #OpenAIRedTeam
-
@agiquality
Agi
on x
OpenAI's red team broke ChatGPT, exposing potential dangers of powerful AI systems. Findings led to improved safety measures before the public release of GPT-4. #AI #ChatGPT #OpenAI https://www.ft.com/...
-
@muradahmed
Murad Ahmed
on x
Superb reporting - @madhumita29 spoke to OpenAI's “red team” - the experts hired to try and “break” GPT4. Amongst other things, one chemical professor used it to create a nerve agent - in @FT https://www.ft.com/...
-
@financialtimes
@financialtimes
on x
We talked to the eclectic mix of people inside OpenAI's ‘red team’ — experts hired to ‘adversarially test’ GPT-4 with probing or dangerous questions https://www.ft.com/...