Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

What is AI, what is it not, how we use it in physics and how it impacts... you

Claire David

TL;DR

The paper addresses the need to critically examine AI in physics by defining AI, contrasting it with explicit programming, and surveying its use in high-energy physics (HEP). It emphasizes a data-centric paradigm, highlights emerging approaches such as Simulation-Based Inference and uncertainty-aware methods, and discusses societal harms and the necessity for responsible adoption. The work argues that AI should augment—not replace—physics insight, and calls for institutional and individual practices that improve explainability, metric development, and interdisciplinary collaboration. Ultimately, it stresses equipping physicists to navigate the AI revolution as both researchers and informed citizens, with a focus on preserving foundational understanding while leveraging advanced tools.

Abstract

Artificial Intelligence (AI) and Machine Learning (ML) have been prevalent in particle physics for over three decades, shaping many aspects of High Energy Physics (HEP) analyses. As AI's influence grows, it is essential for physicists $\unicode{x2013}$ as both researchers and informed citizens $\unicode{x2013}$ to critically examine its foundations, misconceptions, and impact. This paper explores AI definitions, examines how ML differs from traditional programming, and provides a brief review of AI/ML applications in HEP, highlighting promising trends such as Simulation-Based Inference, uncertainty-aware machine learning, and Fast ML for anomaly detection. Beyond physics, it also addresses the broader societal harms of AI systems, underscoring the need for responsible engagement. Finally, it stresses the importance of adapting research practices to an evolving AI landscape, ensuring that physicists not only benefit from the latest tools but also remain at the forefront of innovation.

What is AI, what is it not, how we use it in physics and how it impacts... you

TL;DR

The paper addresses the need to critically examine AI in physics by defining AI, contrasting it with explicit programming, and surveying its use in high-energy physics (HEP). It emphasizes a data-centric paradigm, highlights emerging approaches such as Simulation-Based Inference and uncertainty-aware methods, and discusses societal harms and the necessity for responsible adoption. The work argues that AI should augment—not replace—physics insight, and calls for institutional and individual practices that improve explainability, metric development, and interdisciplinary collaboration. Ultimately, it stresses equipping physicists to navigate the AI revolution as both researchers and informed citizens, with a focus on preserving foundational understanding while leveraging advanced tools.

Abstract

Artificial Intelligence (AI) and Machine Learning (ML) have been prevalent in particle physics for over three decades, shaping many aspects of High Energy Physics (HEP) analyses. As AI's influence grows, it is essential for physicists as both researchers and informed citizens to critically examine its foundations, misconceptions, and impact. This paper explores AI definitions, examines how ML differs from traditional programming, and provides a brief review of AI/ML applications in HEP, highlighting promising trends such as Simulation-Based Inference, uncertainty-aware machine learning, and Fast ML for anomaly detection. Beyond physics, it also addresses the broader societal harms of AI systems, underscoring the need for responsible engagement. Finally, it stresses the importance of adapting research practices to an evolving AI landscape, ensuring that physicists not only benefit from the latest tools but also remain at the forefront of innovation.

Paper Structure

This paper contains 13 sections, 2 equations, 3 figures.

Table of Contents

  1. Introduction
  2. What is AI?
  3. A few basic questions
  4. The paradigm shift from explicit programming
  5. The core mathematics of AI/ML
  6. AI: What it is not
  7. AI in Physics
  8. Brief overview of AI in HEP
  9. The latest trends
  10. A few questions to physicists and humans
  11. AI and You
  12. Ground rules
  13. Takeaways

Figures (3)

  • Figure 1: The fundamental change from explicit programming (top in blue) versus machine learning (upper green, supervised machine learning, lower green, unsupervised machine learning). Illustration from the author.
  • Figure 2: t-SNE projection of abstract embeddings with K-means clustering for ATLAS Workshop notes between 2022 and 2024 talk_murnane_ml_hep.
  • Figure 4: Illustrations highlighting various negative impacts of AI in society. Top left: The Social Dilemma, Netflix, 2020. Top right: OWL ESG & College of Information and Computer Sciences, University of Massachusetts Amherst owlesg2024. Bottom left: Cambridge Analytica scandal from www.performancemarketingworld.com webarticle_cambridge_analytica. Bottom middle: A viral quote from Joanna Maciejewska maciejewska2024. Bottom right: FactSet and company filings, Karen Weise, The New York Times NYTimes2024.