Your Undivided Attention

The Promise and Peril of Open Source AI with Elizabeth Seger and Jeffrey Ladish

Nov 21, 2023

38:44

Snipd AI

Elizabeth Seger and Jeffrey Ladish, AI development experts, discuss the complexities of open source AI systems, the need for true democratization of AI, the promise and peril of fine-tuning models, and the risks of open source AI. They also explore the importance of regulation and policy in AI development.

AI Summary

Highlights

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

Open sourcing AI models can accelerate scientific research and democratize access, but caution is needed to mitigate potential misuse and ensure responsible decision-making.

Clear definitions of AI democratization and caution in handling highly capable models are necessary to avoid 'democracy washing' and assess risks effectively.

Deep dives

The Debate Over Open Sourcing AI

The podcast explores the debate over open sourcing artificial intelligence (AI) and its potential consequences. Opening up AI models could accelerate scientific research, boost productivity, and democratize access to AI technology. However, it also raises concerns about potential misuse, such as the creation of biological weapons or disinformation campaigns. The irreversible nature of open source AI models amplifies the importance of making the right decisions. The podcast emphasizes the need for clear definitions, rigorous risk assessments, and responsible regulation by governments. It suggests alternatives to full open sourcing and highlights the importance of international coordination in shaping a stable state for AI development.

Democratizing Use

00:50

Introduction

2min

The Complexities of Open Source AI Systems

4min

Democratizing AI: Exploring Multiple Meanings

16min

The Promise and Peril of Fine-Tuning Models

2min

Regulating the Future of AI

9min

The Promise and Peril of Open Source AI

3min

As AI development races forward, a fierce debate has emerged over open source AI models. So what does it mean to open-source AI? Are we opening Pandora’s box of catastrophic risks? Or is open-sourcing AI the only way we can democratize its benefits and dilute the power of big tech?

Correction: When discussing the large language model Bloom, Elizabeth said it functions in 26 different languages. Bloom is actually able to generate text in 46 natural languages and 13 programming languages - and more are in the works.

RECOMMENDED MEDIA

Open-Sourcing Highly Capable Foundation Models

This report, co-authored by Elizabeth Seger, attempts to clarify open-source terminology and to offer a thorough analysis of risks and benefits from open-sourcing AI

BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13B

This paper, co-authored by Jeffrey Ladish, demonstrates that it’s possible to effectively undo the safety fine-tuning from Llama 2-Chat 13B with less than $200 while retaining its general capabilities

Centre for the Governance of AI

Supports governments, technology companies, and other key institutions by producing relevant research and guidance around how to respond to the challenges posed by AI

AI: Futures and Responsibility (AI:FAR)

Aims to shape the long-term impacts of AI in ways that are safe and beneficial for humanity

Palisade Research

Studies the offensive capabilities of AI systems today to better understand the risk of losing control to AI systems forever

RECOMMENDED YUA EPISODES

A First Step Toward AI Regulation with Tom Wheeler

No One is Immune to AI Harms with Dr. Joy Buolamwini

Mustafa Suleyman Says We Need to Contain AI. How Do We Do It?

The AI Dilemma

Your Undivided Attention is produced by the Center for Humane Technology. Follow us on Twitter: @HumaneTech_