The AI Canvas - Navigating the Complexities of AI: A conversation with Chris Stokel-Walker

In this captivating episode of the AI Canvas podcast, host David Foster engages in a compelling conversation with technology journalist Chris Stokel-Walker. Together, they delve into a wide range of topics related to AI and its impact on various aspects of society.

Chris discusses the concerns surrounding the use of ChatGPT and the potential risk of incorrect information. He also highlights the limitations of information summarisation in journalism. Furthermore, they explore the role of AI tools like ChatGPT as a starting point for writing and the challenges of aligning large language models with truth.

The conversation takes a deep dive into the battle between genuine journalism and deception, emphasising the impact of misinformation and the challenges faced by journalists in debunking fake news.

Join us for this enlightening episode as we explore the evolving landscape of AI and its profound implications. Tune in now to gain valuable insights and perspectives from Chris, a leading voice in the field of technology journalism

15:47 - Chris's concern about using ChatGPT and the risk of having incorrect information
19:16 - The use of ChatGPT for information summarization and the limitations in journalism
23:22 - Chris's discussion on using ChatGPT as a starting point for writing and filling in details
27:28 - The concern about potential societal laziness and outsourcing of thinking with the rise of AI tools
29:13 - The concern about misinformation and the battle between genuine journalism and deception
32:54 - The impact of misinformation and the challenges faced by journalists in debunking fake news
34:27 - The potential futility of fighting back against misinformation due to its widespread proliferation
36:41 - The discussion on aligning large language models to truth and the challenges of defining "truth"
52:54 - The impact of language and the reverse causality between language and intelligence
56:42 - The mystical concept of language models exhibiting emerging capabilities
58:47 - The idea of autoregressive models and next token prediction
01:00:25 - Speculating on the future of AI in the next five years
01:03:34 - Competition among big tech platforms in developing the best LLM
01:04:00 - The skepticism around Apple glasses and the concept of living in a metaverse
01:06:34 - The need for regulation in the AI industry and the challenges of government regulation
01:09:52 - The potential challenges of AI regulation keeping up with the rapid pace of AI development

Connect with Chris Stokel-Walker

Connect with David Foster

Follow ADSP on LinkedIn for more AI insights and discussions.

ADSP is an innovative consultancy dedicated to delivering value through the application of Data Science, AI and Generative AI. We build bespoke solutions that enable people and companies to thrive.

Check out our website here

For access to the condensed notes and full transcript, including key quotes and takeaways please get in touch at hello@adsp.ai

Subscribe to the AI Canvas podcast for more exciting episodes exploring the intersection of AI and various industries.