Artificial intelligence definitions

Negotiating Tech: An Inventory of U.S. Union Contract Provisions for the Digital Age

Technology definitions


1.2 Artificial intelligence definitions

Given the rapid evolution and broad applicability of artificial intelligence (AI) in the workplace, many union contracts include specific definitions to clarify the scope of covered technologies. These definitions often distinguish between general AI, machine learning, generative AI, and traditional rule-based systems. While terminology may vary across agreements, the goal is to establish clear boundaries for how AI-related provisions apply, especially as new tools emerge that blend predictive analytics, content generation, and automated decision-making.

Examples: 

  • Artificial Intelligence (AI) is a catch-all term that refers broadly to a machine-based system’s ability to analyze, apply logic, and improve its capabilities through data analysis to solve tasks. An “AI System” is any machine-based system that relies on AI as a primary function. “Machine Learning (ML)” is a subset of AI that enables machines to create algorithms based on data patterns (e.g., diffusion models, large language models) to perform specific tasks like predicting behavior or generating content. “Generative AI (Gen AI)” is a subset of ML focused on producing new content (e.g., text, video, audio, 3D models). Gen AI systems are defined by their content-generating abilities. “Deep Learning” is a deeper subset of ML utilizing artificial neural networks with multiple layers to process data. The terms “Gen AI” and “Deep Learning” are used here for simplicity, and this provision will apply to any technology that aligns with these definitions, regardless of its specific name or label. (IATSE-PCBA46)
  • “The parties acknowledge that definitions of Generative Artificial Intelligence (‘GAI’) vary, but agree that the term generally refers to a subset of artificial intelligence that learns patterns from data and produces content based on those patterns (e.g., ChatGPT4, MidJourney, Dall-E2). It does not include ‘traditional AI’ technologies programmed to perform specific functions (e.g., CGI and VFX), such as those already used during all stages of motion picture production (e.g., pre-visualization, preproduction, production, post-production, distribution, marketing). The term GAI is used for convenience and this Section [XX.1] shall also apply to any technology that is consistent with the foregoing definition, regardless of its name.” (SAG-AFTRA-PCBA2)
  • “The parties acknowledge that definitions of generative artificial intelligence (‘GAI’) vary, but agree that the term generally refers to a subset of artificial intelligence that learns patterns from data and produces content, including written material, based on those patterns, and may employ algorithmic methods (e.g., ChatGPT, Llama, MidJourney, Dall-E). It does not include ‘traditional AI’ technologies such as those used in CGI and VFX and those programmed to perform operational and analytical functions.” (WGA-PCBA1)
  • “AI is defined as a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. AI systems use machine and human-based inputs to perceive real and virtual environments; abstract such perceptions into models through analysis in an automated manner; and use model inference to formulate options for information and action.” (IAMAW-PCBA54)
  • “Artificial intelligence is the simulation of human intelligence processes by machines, especially computer systems, to perform tasks requiring cognition, including, but not limited to, learning, reasoning, self-correction, and creativity.” (PLEA-PCBA110)
  • “Generative Artificial intelligence (“Generative AI”) includes generative systems of technology that are capable of creating content, in whole or in part, through the utilization of machine learning techniques, including but not limited to generative adversarial networks, generative pre-trained transformers (e.g., ChatGPT), and variational autoencoders. As defined in this section, Generative AI does not include any existing technologies used to assist in the creation of content such as a spell-check, autocorrect, grammar checks, etc.” (ZDCG-TNG-CWA-PCBA111)