Published May 1, 2019

Maintaining Human Control of Artificial Intelligence with Joanna Bryson - TWiML Talk #259

Joanna Bryson explores the critical need for human control over AI through responsible design, ethical frameworks, and robust governance, highlighting how DevOps principles and regulations can align AI development with societal values and ensure safety.
Episode Highlights
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) logo

Popular Clips

Episode Highlights

  • Natural Intelligence

    Understanding natural intelligence is crucial for advancing AI, yet many leaders in AI lack this knowledge. argues that while some AI experts are ignorant of natural intelligence, their contributions remain significant. She emphasizes the importance of engaging with government and society to address AI's societal impacts 1. Bryson also highlights the complexity of learning architectures, noting that different brain regions solve distinct problems, which informs AI development 2.

    There's never going to be an algorithm that gives you omniscience. There's never going to be a single solution.

    ---

       

    AI Safety

    Misconceptions about AI safety often overlook the extensive work already being done in the field. points out that AI is pervasive in software, with thousands working on safety and security, contrary to claims of limited efforts 3. She also discusses the challenges in AI regulation, highlighting the European Union's efforts and the need for better understanding of AI as a product 4.

    Everybody's using AI, everybody's using machine learning, which is a huge change.

    ---