Vector Institute launches UnBias-Plus bias-detection toolkit

The Vector Institute released UnBias-Plus on June 30, 2026, a free toolkit that detects, explains, and rewrites biased language in written content and AI training datasets. The toolkit features multi-class bias classification, biased-span localization, neutral-text rewriting, and per-decision reasoning, with code and models publicly available.

For practitioners: Tools that automate bias detection and neutral rewriting can materially change dataset curation, content review, and compliance workflows. The Vector Institute released UnBias-Plus on June 30, 2026, a free toolkit that Vector researchers describe as able to detect, explain, and rewrite biased language in written content and AI training datasets, according to a GlobeNewswire press release and BetaKit reporting. An arXiv preprint for the project documents features including segment-level multi-class bias classification, biased-span localization, neutral-text rewriting, and per-decision reasoning, and lists code and models as publicly available; the preprint also notes code metadata release v0.1.6 and runtime requirements Python =3.10 , GPU with CUDA 12.4 recommended . BetaKit quotes Vector applied-ML scientist Shaina Raza on the project rationale.