Hardware Podcasts

Building a Framework to Trust AI

EE Times Current · 24 Sept 2021 · 29m

What this episode covers

Helen Toner, Director of Strategy at Georgetown University’s Center for Security and Emerging Technology, argues that AI currently lacks the engineering discipline required for mission-critical systems, describing current machine learning development as akin to 'alchemy.' She outlines a necessary framework involving basic science, socio-legal standards, and engineering rigor to ensure reliability, citing failures in social media algorithms and hiring tools as evidence of current risks. The episode also covers industry news, including Arm’s Sophie architecture for automotive software and Silicon Labs’ new in-line programming capabilities for IoT security.

Most relevant to

early-stage companies

Topics

Mentioned

Companies
Georgetown University, Center for Security and Emerging Technology, Arm, Silicon Labs, NIST, NTSB
Products
Sophie
Also
Matt Johnson

Related episodes from other shows

Other shows on the same ground, closest first.

  • 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

  • Embedded.fm · 1h 16m

    421: Paint the Iceberg Yellow

    Chris Hobbs, who works in kernel development at BlackBerry QNX and authored 'Embedded Software Development for Safety-Critical Systems,' explains the distinction between safety-critical and standard embedded systems, emphasizing safety culture and the concept of SOTIF (Safety of the Intended Functionality). He details how modern systems, such as autonomous vehicles and medical devices, face risks from 'accidental systems' where interactions between components create unforeseen hazards, citing examples like a radar system dependent on jammed GPS. Hobbs advocates for goal-based standards like UL 4600 and the use of Safety Cases with Goal Structuring Notation to demonstrate safety rather than relying on prescriptive checklists. He also discusses the challenges of verifying machine learning algorithms, the limitations of languages like C and Rust in safety contexts, and the importance of structured risk analysis using tools like Bayesian fault trees and LaTeX for traceability.

    Embedded firmwareSecurity & privacy
    12 chapters

  • Embedded.fm · 1h 10m

    514: Just Turn Off All the Computers

    Philip Koopman, a retired Carnegie Mellon University professor and author of 'Embodied AI Safety: Reimagining safety engineering for artificial intelligence in physical systems,' discusses the critical safety, security, and legal challenges of deploying AI in physical devices. He uses Waymo robotaxis as a primary example to illustrate how 'moral crumple zones' and human factors create liability gaps when AI supplants human agency. Koopman argues that current tort negligence laws are insufficient for holding manufacturers accountable for AI errors, necessitating a shift in how engineers approach hazard assessment and risk mitigation for embodied systems.

    AI & machine learningSecurity & privacy
    8 chapters

  • Global Medical Device Podcast · 53m

    #464: Why LLM Unpredictability is a Liability in MedTech

    Tyler Harmon, CEO of Iaso Automated Medical Systems, explains why the inherent randomness in Large Language Models (LLMs) makes them unsuitable for safety-critical medical devices compared to deterministic algorithms like Convolutional Neural Networks (CNNs). He details how Iaso uses non-LLM machine learning to detect Acute Respiratory Distress Syndrome (ARDS) in ICU patients and advocates for treating third-party AI stacks as Software of Unknown Provenance (SOUP). Harmon also outlines the need for a dedicated Machine Learning Management System (AIMS) within Quality Management Systems to govern AI governance, hallucination risks, and data integrity.

    AI & machine learningQuality & yield
    7 chapters

  • Embedded.fm · 1h 05m

    437: Chirping With the Experts

    Daniel Situnayake, Head of Machine Learning at Edge Impulse and co-author of the O'Reilly book 'AI at the Edge', discusses the practical application of embedded machine learning. He explains the BLERP framework (Bandwidth, Latency, Economics, Reliability, Privacy) for deciding when to deploy AI on edge devices and emphasizes the importance of responsible AI, including hiring ethicists and using diverse datasets. The conversation also covers the current state of large language models like ChatGPT, noting their utility as creative tools with human-in-the-loop oversight rather than autonomous agents.

    AI & machine learningEmbedded firmware
    10 chapters

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