Generative AI: Oh God What Now?: Live from SXSW 2023

·34 min·2 clips

Samir says early AI models have accuracy problems, making healthcare use risky for manufacturers.

1. The Real Chemistry Podcast episode 'Generative AI, Oh God, What Now?' was recorded live at South by Southwest in Austin, Texas. 2. Aaron Strout is Real Chemistry's Chief Marketing Officer and host, Romain Bogarts is Director of Artificial Intelligence Product Management, and Samir Grover is an AbbVie commercial healthcare leader. 3. The episode asks whether generative AI is a fad or an 'iPhone moment' for healthcare and pharmaceutical work. 4. Romain defines AI as a broad field combining computer science, big data, mathematics, and statistics to make predictions. 5. He says machine learning uses historical data, while generative AI uses deep learning and large language models to generate new content. 6. He points to ChatGPT and DALL-E as examples of systems that can produce text and images. 7. Romain says the general public is seeing and interacting with AI for the first time, which makes the moment different from earlier research phases. 8. Samir says he has spent more than a decade in pharmaceuticals, launching drugs in cardiovascular, gastroenterology, psychiatry, and neuroscience. 9. He also says he has flown for seven years, owns a plane with two partners, and treats flying from New Jersey to Boston as 'like a time machine.' 10. The conversation returns to healthcare adoption, where both speakers say companies should solve business problems first and use AI to support patients and healthcare providers. 11. Samir says early generative AI models raise accuracy concerns, especially in a 'risk adverse' environment like a large manufacturer. 12. He warns that 'garbage in, garbage out' applies to healthcare data, including demographics and patient populations. 13. He says generative AI can use both structured and unstructured data, including PowerPoint material and clinical-trial information. 14. Samir says generative AI lacks common sense and gives examples involving driving in Texas versus New York and simple toe-counting logic. 15. He also raises ethical and legal questions about who gets paid when AI tools generate copy for brand.com or CRM content. 16. Romain adds that transparency, compliance, and a 'taught approach' to legal components matter when choosing vendors and partners. 17. The episode has a conversational interview style, with Aaron and Romain pushing Samir on definitions, risks, and practical use cases. 18. The energy stays upbeat and exploratory, with repeated back-and-forth on augmentation, process redesign, and feedback loops. 19. Best for healthcare marketers, pharma leaders, and AI strategy teams. 20. Probably skip if you want a purely technical AI research discussion.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from The Real Chemistry Podcast