About The CCAI Glossary

The CCAI Glossary is a set of cross-industry definitions developed and ratified by the people who make film, television, and content.

Our goal is to create a shared language for how we, as an industry, discuss, and make decisions about how (and if) we want to use AI in our work.

It is a living document that will be updated as the technology evolves, but we need your help! Each definition has a field to invite your feedback. If you are interested in becoming a Definitions Delegate, please sign up by clicking “Get Involved” on the main page.

Our Goal

Generative AI is here, and its impact is being felt unevenly across creative work. Without shared terms and comparable evidence, people are left to make sense of a new, complex technology directly impacting their industry all from headlines, vendor claims or isolated experience.

CCAI gives the creative community common reference points: definitions discussed and ratified across disciplines, derived from real-world, role-level lived experience and evidence about how AI technologies and capabilities modify work, and a way for creatives to test that evidence against what they are seeing themselves.

The goal is not to predict one future for the entire industry. It is to make complex, ongoing change understandable and visible. We can upgrade our industry's systems and institutions. When people can see what colleagues in their own roles are reporting — using (or not using) AI tools, discussing new technology and its capabilities and impacts, and identifying where education and cross-skilling may help — they are able to respond with clarity and agency. As further testimonies come in, the evidence becomes clearer and more useful to creatives, helping to turn fear and confusion into empowerment.

About CCAI

The Creators Coalition on AI is a tech-agnostic convening organization created as a central hub for cross-industry discussion about how AI is affecting entertainment and creative communities.

The wider coalition includes members of the DGA, SAG-AFTRA, WGA, Producers United, PGA, and IATSE, alongside independent artists, executives, and technologists. People participate as individuals in an open discussion—not as organizational endorsements. To see our founders and signatories, click here.

MODEL WEIGHTS

Numerical parameters in an AI model that capture statistical patterns derived from its training process. These weights shape how the model processes inputs and generates outputs.

History & Rationale

Pre-Publication V3. Numerical parameters derived from a data point within a dataset which are assigned to components of an AI model adjusted during training. Model weights to help influence how the model processes inputs and generates outputs by representing statistical relationships from training data.

‍

Deliberation Summary:

Language edits for clarity.

‍

Pre-Publication V2. A numerical parameter derived from a data point within a dataset assigned to components of within an AI model adjusted during training to help influence how the model processes that helps determine the model’s overall behavior, memory and outputs in response to inputs and generates outputs by representing statistical relationships from training data.

‍

Deliberation Summary:

Not readily understandable for a reader how a numerical parameter can determine a model's behavior, and this needed more context on that point, including that the parameters are derived from points within the dataset.

‍

Pre-Publication V1. A numerical parameter within an AI model that helps determine the model’s overall behavior, memory and outputs in response to inputs.

‍

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form. Please refresh/reload the page and try again.

MODEL WEIGHTS

Have Improvements to this definition?