The Vector Institute has launched a free, open-source AI tool designed to help organizations detect, explain, and rewrite biased language in written content and AI training data.
Called UnBias-Plus, the tool was developed by Vector’s AI Safety research scientists and is aimed at professionals working in areas where language can have real-world consequences, including human resources, workplace equity, health care, insurance, media, and AI development.
The Toronto-based institute says UnBias-Plus can analyze text for biased language across dimensions including race, gender, age, and political framing. The tool explains why specific language has been flagged, suggests more neutral alternatives, and returns a rewritten version of the original content.
For HR, EDI, and workplace teams, Vector says the tool can be used to review job descriptions, performance reviews, promotion cases, and internal communications for wording that may unintentionally create barriers for candidates or employees.
The launch comes as organizations place greater scrutiny on how bias enters both workplace processes and AI systems. Vector noted that bias in AI training data has been measured at rates ranging from 3.4 per cent to 38.5 per cent, while even safety-tuned large language models can exhibit implicit racial and gender bias.
“What drove us to build this was simple: the people most harmed by biased language are often the last to know it’s there,” said Shaina Raza, PhD Applied Machine Learning Scientist, Responsible AI, at Vector Institute. “A job candidate doesn’t know why a door keeps closing. We wanted to make the invisible visible — a free tool that surfaces bias, explains it and helps fix it.”
Raza said the tool is also intended for AI developers and data scientists who want to identify and clean biased language before it becomes part of a model’s training data.
“Bias doesn’t just live in a document someone writes today,” she said. “It lives in the data used to train AI models and, when it goes undetected there, it gets learned, replicated, and deployed at scale across entire organizations.”
UnBias-Plus is available in two formats. A browser-based version supports content of up to 750 words and is designed for public use and organizational testing. A developer installer is also available for teams that want to integrate the tool into internal applications, systems, and workflows.
For talent teams, the release reflects a growing shift in how AI is being applied inside organizations: not only to automate work, but to improve decision-making, communication, and fairness in hiring and employee management.
Vector says the tool can be used by editors to review stories before publication, by health care teams to assess clinical notes and patient communications, by insurers to review coverage-related language, and by AI teams to screen prompts, annotations, source text, and model outputs.
“Canada’s AI leadership depends on building systems that create more equitable outcomes — and that means ensuring the research behind those systems is accessible to everyone,” said Kathryn Hume, VP AI Engineering at Vector Institute. “Making UnBias-Plus free, transparent and open-source is a deliberate expression of Vector’s mission as an independent, not-for-profit institute.”
UnBias-Plus is the first release in what Vector says will become a broader suite of Safe AI tools. Future updates are expected to include audio and video bias detection, as well as misinformation detection tools.