Hardware Podcasts

112. Overcoming Risks of AI Misuse

The Next Byte Podcast · 7 Mar 2023 · 27m

What this episode covers

Hosts Daniel and Forbode discuss research from the University of Michigan and Copenhagen Business School, led by Professors Nigel Melville, Lionel Robert, and Xiao Xiao, on mitigating AI risks by keeping humans in the loop. They analyze four critical areas: decision making, creativity automation, machine-human relationships, and machines teaming with machines, using examples like Tesla Autopilot, DALL-E, and deep fakes. The episode emphasizes that while AI offers productivity gains, unsupervised deployment in high-stakes domains like hiring or autonomous driving requires human governance to prevent bias and harm. The hosts argue for a balanced approach that leverages AI as a tool for collaborative intelligence rather than replacing human oversight entirely.

Most relevant to

large enterprises

Topics

Mentioned

Companies
University of Michigan, Copenhagen Business School, Tesla, Wevolver
Products
DALL-E, Stable Diffusion, GPT
Also
Nigel Melville, Lionel Robert, Xiao Xiao, Missy Cummings

Related episodes from other shows

Other shows on the same ground, closest first.

  • The Builder Circle · 1h 13m

    S3 E5: AI in Hardware: The Future of Design, Manufacturing, and Innovation with Kristen Edwards

    Kristen Edwards, CTO and co-founder of Pensa Technologies and MIT PhD candidate, discusses integrating AI into hardware engineering workflows, emphasizing the need to track decision reasoning and manage data governance. She explains that while AI acts as a 'co-pilot' to augment human expertise, high-consequence tasks require human oversight due to liability and explainability concerns. Edwards highlights specific challenges in spatial reasoning, design-for-manufacturing translation, and IP security, suggesting techniques like federated learning and retrieval-augmented generation to protect sensitive engineering data. The conversation covers practical applications in CAD, automated testing, and quality assurance, noting that AI will likely replace repetitive tasks rather than creative engineering roles.

    AI & machine learningSecurity & privacy
    9 chapters

  • How I Built This · 28m

    The peril (and promise) of AI with Tristan Harris: Part 1

    Tristan Harris, co-founder of the Center for Humane Technology, explains how the incentive structures of the attention economy drive harmful outcomes, a lesson he first learned while building his company Apture in the mid-2000s. He details his 2013 internal Google presentation warning about the impact of designers at Google, Apple, and Facebook on global attention, and how the rapid adoption of social media and now AI outpaces human cognitive adaptation. Harris argues that the race to build artificial general intelligence by companies like OpenAI, Anthropic, and DeepMind mirrors the Manhattan Project, creating existential risks through uncontrolled scaling of Transformer-based models.

    AI & machine learningStarting a product company
    6 chapters

  • The Pragmatic Embedded Podcast · 55m

    Crossover with Embedded AI Podcast

    Jeff Gable, Luca, and Ryan Torvik discuss integrating Large Language Models (LLMs) into embedded systems engineering workflows, emphasizing that AI must augment rather than replace rigorous engineering discipline. They identify 'vibe coding' as a critical anti-pattern where developers lose control of architectural decisions by letting AI generate too much code at once, leading to unmanageable outputs and hallucinations. The hosts advocate for strict context management, test-driven development (TDD), and small, incremental commits to maintain deterministic quality, particularly in safety-critical domains like medical devices and automotive systems. They conclude that while LLMs accelerate prototyping and boilerplate generation, human code review and ownership remain non-negotiable for reliable production software.

    Embedded firmwareDebugging & failure analysis
    8 chapters

  • Built 2 Scale · 55m

    The Value of Human Cognition Is Going Negative. Build the Outcome Machine.

    Scott Wilcox and Matt Perrott of BuildPass discuss the shift from selling outputs to building 'outcome machines' that encode domain expertise into AI-native systems. They highlight the risks of unmanaged AI spending, citing a specific incident where an OpenRouter bill exploded due to uncapped API access, and discuss how Tesla's Cybercab and Model Y illustrate the convergence of software and hardware manufacturing. The conversation explores computational design, self-healing physical products, and the labor shortage in skilled trades, suggesting that companies must harness frontier technology to deliver reliable outcomes rather than relying on fragmented human cognition.

    Connected products & IoTAI & machine learning
    9 chapters

  • How I Built This · 32m

    The peril (and promise) of AI with Tristan Harris: Part 2

    Tristan Harris, co-founder of the Center for Humane Technology, discusses the existential risks of generative AI, particularly the collapse of trust in digital evidence and the potential for deepfakes to undermine legal and social systems. He argues that current incentives drive a reckless race for capabilities and advocates for liability frameworks and authenticated messaging standards to slow development and ensure safety. Harris compares the current AI landscape to the nuclear age, highlighting the need for global cooperation, supply chain controls on advanced chips, and a shift from 'race to power' to 'race to safety'.

    AI & machine learningPolicy, trade & tariffs
    7 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Next Byte Podcast. The episode itself belongs to its publisher.