Understanding AI & Bargaining: An Informational Toolkit

Cette ressource est également disponible en français.

The collective agreements for CUPE 3903 Units 1, 2, & 3 expire on August 31^st, 2026. This means that as soon as June, we can begin negotiating new contracts. Right now, we have an opportunity to address the ways in which Artificial Intelligence (AI) impacts our working conditions.

AI is not just one thing. It is a broad term for computer systems that perform tasks that usually involve human judgment, pattern recognition, prediction, or language processing. Many (but not all) workplace tools being introduced right now are “narrow AI”, meaning they do specific tasks rather than “think” like a person.

Different Types of AI

• Artificial Intelligence (AI): The broad umbrella term for systems that perform tasks associated with human intelligence, such as recognizing patterns, making predictions, or processing language. 

• Machine Learning (ML): A subset of AI that learns from data and improves pattern recognition or prediction without being explicitly programmed for every step. 

• Deep Learning: A subset of machine learning that uses large neural networks and is common in image recognition, speech systems, and generative tools. 

• Generative AI: Systems that produce new content such as text, images, code, audio, or summaries based on patterns in training data. Also referred to as strong AI.  

• Large Language Models (LLMs): A type of generative AI trained on very large text datasets and used for chatbots, drafting, summarizing, translation, and coding assistance. 

• Computer Vision: AI that interprets images or video, such as facial recognition, object detection, or visual surveillance systems. 

• Natural Language Processing (NLP): AI focused on understanding, analyzing, or generating human language; many LLM tools are part of this area. 

• Predictive AI: Systems that forecast outcomes, rank people, or score risk, often used in hiring, scheduling, performance management, or student analytics. 

• Surveillance or algorithmic-management AI: Systems that monitor workers through keystrokes, screen activity, location, biometrics, or productivity scoring. 

AI at Work

To figure out how AI might be impacting your work at York, you can start with asking these three simple questions about any AI tool:

1. What data does it use?  

2. What decision does it influence?  

3. Who is accountable?  

Various labour resources about digital-age bargaining emphasize that understanding AI isn’t just about technical knowledge. Understanding AI includes understanding surveillance, data extraction, bias, de-skilling, and workload intensification.

For more useful references, you can click here to check out CUPE’s guide to artificial intelligence, and you can click here to see CUPE’s guide to bargaining strong collective agreements for the digital age.

For workplace discussions, it helps to sort tools into practical categories:

• Content tools: Chatbots, text generators, image generators, coding assistants. 

• Decision tools: Ranking, scoring, prediction, recommendation, and risk assessment systems. 

• Monitoring tools: Facial recognition, keystroke logging, screen capture, voice analysis, and automated productivity tracking. 

• Administrative tools: Scheduling systems, workflow allocation, and automated support portals that may quietly shape workload and evaluation. 

Questions members should ask:

• What kind of AI is this, and what exactly does it do? 

• Is it generating content, making predictions, or monitoring people? 

• What data is being collected, where does that data go, and what happens to the data? 

• Does it affect workload, evaluation, discipline, hiring, or job security? 

• Can a human review, override, or contest its outputs? 

• Is anyone’s teaching material, research, student work, or personal data being used to train it? 

Why It Matters

At York University, like elsewhere in the sector, we’re seeing Artificial Intelligence becoming more and more part of our workplace all the time. As CUPE 3903 members, we need collective agreements that keep up with this changing reality and put workers first.

Understanding AI better can help us recognize the ways in which workplace systems are not neutral “efficiency tools”, but sometimes act as forms of management, surveillance, or labour restructuring.

Bargaining for AI protections means centering our humanity in our work as educators and academic workers. Whether you’re a Teaching Assistant, Contract Faculty, or RA/GA, we all deserve accountability, human oversight, job protection, data privacy, intellectual property rights, transparency, and the ability to challenge automated decisions.