meta_pixel
Tapesearch Logo
Log in
Science Magazine Podcast

FDA clinical trial protection failures, and an AI that can beat curling’s top players

Science Magazine Podcast

Science Podcast

News, News Commentary, Science

4.3842 Ratings

🗓️ 1 October 2020

⏱️ 31 minutes

🧾️ Download transcript

Summary

Investigative journalist Charles Piller joins host Sarah Crespi to discuss his latest Science exclusive: a deep dive into the Food and Drug Administration’s protection of human subjects in clinical trials. Based on months of data analysis and interviews, he uncovered long-term failures in safety enforcement in clinical trials and potential problems with trial data used to make decisions about drug and device approvals. Sarah also talks with Klaus-Robert Müller, a professor of machine learning at the Technical University of Berlin, about an artificial intelligence (AI) trained in the sport of curling—often described as a cross between bowling and chess. Although AI has succeeded in chess, Go, and poker, the constantly changing environment of curling is far harder for a nonhuman mind to adapt to. But AIs were the big winners in competitions with top human players, Müller and colleagues report this week in Science Robotics. This week’s episode was produced with help from Podigy. Listen to previous podcasts. About the Science Podcast Download a transcript (PDF).    ++Meta   [Image: Cory Denton/Flickr; Music: Jeffrey Cook]   Authors: Sarah Crespi; Charles Piller   Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript

Click on a timestamp to play from that location

0:00.0

This podcast is supported by the Icon School of Medicine at Mount Sinai, one of America's leading research medical schools.

0:07.8

Icon Mount Sinai is the academic arm of the eight hospital Mount Sinai health system in New York City.

0:13.9

It's consistently among the top recipients of NIH funding.

0:18.0

Researchers at Icon Mount Sinai have made breakthrough discoveries in many fields vital to

0:23.0

advancing the health of patients, including cancer, COVID and long COVID, cardiology, neuroscience, and

0:30.4

artificial intelligence. The Icon School of Medicine at Mount Sinai. We find a way.

0:48.7

Welcome to the science podcast for October 2nd, 2020. I'm Sarah Presby. Each week, we feature the most interesting news and research published in science and the sister journals. First up,

0:53.4

we have investigative journalist

0:54.6

Charles Piller. He talks about his latest story on long-term failures at the FDA to regulate

1:00.6

clinical trial safety and data integrity. We also hear from machine learning researcher

1:06.0

Klaus Robert Mueller about an AI that can beat human experts in the sport of curling. Now we have contributing

1:15.4

correspondent Charles Pillar. He wrote an investigative feature on FDA's serious decline in

1:21.0

protections for patients and for data integrity. Hi, Charlie. Hey, Sarah. I'm not sure everyone knows that FDA has oversight of

1:29.9

clinical trials. What exactly is this agency's role with respect to human trials? The FDA is a

1:37.1

vast agency with many responsibilities, and in this case, it has two elements of control over the quality of clinical trials.

1:47.9

One is whether patients are properly protected during the trials and also whether the data

1:54.7

generated from these trials is reliable.

1:57.6

And in keeping with the important goal of the FDA of approving drugs and

2:04.6

devices that actually were.

2:07.1

But as you found in your investigation, the FDA enforcement has been light-handed, slow-moving,

2:13.9

and secretive. Can you give us some examples that you found of shoddy clinical research

2:18.7

practices, poor record keeping that you came across in your deep dive into this?

...

Please login to see the full transcript.

Disclaimer: The podcast and artwork embedded on this page are from Science Podcast, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of Science Podcast and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.

Copyright © Tapesearch 2026.