AI Inside

Latest Episodes:

Follow the Funding of AI Doom Feb 07, 2024

Show notes

This week, Jason Howell and Jeff Jarvis welcome Dr. Nirit Weiss-Blatt, author of "The Techlash and Tech Crisis Communication" and the AI Panic newsletter, to discuss how some AI safety organizations use questionable surveys, messaging, and media influence to spread fears about artificial intelligence risks, backed by hundreds of millions in funding from groups with ties to Effective Altruism.

INTERVIEW

Introduction to guest Dr. Nirit Weiss-Blatt
Research on how media coverage of tech shifted from optimistic to pessimistic
AI doom predictions in media
AI researchers predicting human extinction
Criticism of annual AI Impacts survey
Role of organizations like MIRI, FLI, Open Philanthropy in funding AI safety research
Using fear to justify regulations
Need for balanced AI coverage
Potential for backlash against AI safety groups
With influence comes responsibility and scrutiny
The challenge of responsible AI exploration
Need to take concerns seriously and explore responsibly

NEWS BITES

Meta's AI teams are filled with female leadership
Meta embraces labeling of GenAI content
Google's AI Test Kitchen and image effects generator

Learn more about your ad choices. Visit megaphone.fm/adchoices

Schibsted's AI Strategy with Sven Størmer Thaulow Jan 31, 2024

Show notes

In this episode, Sven Størmer Thaulow, EVP/Chief Data and Technology Officer at Schibsted, joins Jason Howell and Jeff Jarvis to discuss the creation of a Large Language Model (LLM) for media in Norway, the importance of aligning AI models with cultural values, and the potential for American media companies to collaborate on creating English language models. The conversation also explores the use of AI in generating article titles, providing a different user interface, and experimenting with an AI agent for tech reviews.

INTERVIEW

Introduction to the guest, Sven Størmer Thaulow, EVP/Chief Data and Technology Officer at Schibsted.
The work Sven and his team are doing at Schibsted.
Background on Schibsted and its success in the news industry.
Schibsted's successful subscriptions in Norway and their high-quality news products.
The creation of an LLM in the Norwegian language and Schibsted's role in it.
Different attitudes towards AI and language models in Norway and the US.
Schibsted's AI strategy and their work with AI in the recommendation system space.
The importance of data and language data for AI and language models.
Efforts to build Norwegian and Northern Germanic language LLMs.
The need to align AI models with cultural values.
Potential for American media companies to collaborate on creating English language models.
Sven's experiment with using their LLM to generate article titles at Schibsted.
The use of AI to provide a different user interface and relationship with content.
Experiment with an AI agent for tech reviews.

NEWS BITES

Google's new Lumiere AI video generator, Video Poet, can create stunning clips by using space and time together.
Microsoft's changes to its AI text-to-image generation tool in response to reports of people using it to create nonconsensual sexual images of Taylor Swift.
OpenAI and Common Sense Media have partnered to address concerns about the impacts of AI on children and teenagers.
Keep It Shot is a tool that allows Mac users to rename, organize, and quickly find their screenshots using AI.

Learn more about your ad choices. Visit megaphone.fm/adchoices

AI Positive Jan 24, 2024

Show notes

On the premiere episode of the AI Inside podcast, hosts Jeff Jarvis and Jason Howell discuss AI copyright issues with Common Crawl Foundation's Rich Skrenta regarding news outlets limiting access to content they publish publicly, impacting the integrity of Common Crawl's internet archive. In recent years, the archive has been used by LLMs as AI training data, and the implications of restricting information have a dramatic impact on the data quality that survives.

INTERVIEW

Introduction and background on AI Inside podcast
Discussion of the recent AI oversight Senate hearing Jeff testified at
Introduction of guest Rich Skrenta from Common Crawl Foundation
Overview of Common Crawl and its goals to archive the open web
Discussion of how Common Crawl data is used to train AI models
News publishers wanting content removed from Common Crawl
Debate around copyright, fair use, and AI’s “right to read”
Mechanics of how Common Crawl works and what it archives
Concerns about restricting AI access to data for training
Risk of regulatory capture and only big companies being able to use AI
Discussion of recent court ruling related to web scraping
Hopes for Common Crawl's growth and evolution

NEWS BITES

Interesting device announcement from CES - Rabbit R1 with Perplexity AI integration
Study on actual risk of AI automating jobs away in the near future

Learn more about your ad choices. Visit megaphone.fm/adchoices

Welcome to AI Inside Jan 17, 2024

Show notes

AI Inside is your new home for keeping up to date on what's new in the world of Artificial Intelligence. Join Jason Howell and Jeff Jarvis every Wednesday as they speak with some of the world's leading minds about how AI is being used, abused, and built into the fabric of our lives. AI Inside will look at the influence of AI in the newsroom, the ongoing uncertainty around AI art and copyright, the scalability of disinformation, supercharging worker productivity, and so much more. And if a groundbreaking service grabs hold of the news cycle, AI Inside will be there to take a closer look. It's all fair game on the AI Inside podcast.

Learn more about your ad choices. Visit megaphone.fm/adchoices