19/05/2026
🪝 If it's not in the standard, it doesn't exist. Time to change that.
Tomorrow, the BIAS Project joins the Innovation Standardisation Workshop organised by Eloquence AI — and we're coming with something specific to say.
Represented by our partner Digiotouch AI, we will be presenting our work to an audience of AI standards experts and representatives from standards development organisations (ISO/IEC SC 42, CEN/CENELEC JTC 21, NIST, ETSI).
Our focus? A gap that's still largely invisible in current AI standardisation efforts:
🔍 Bias in NLP decision systems used in recruitment and employment — not conversational AI, but the text-based tools that quietly shape who gets hired, shortlisted, or screened out.
At BIAS, we develop open-source tools and methodologies to detect and mitigate algorithmic bias in exactly these systems. Two of our most concrete outputs, the Debiaser and the multilingual NLP Demonstrator, are directly relevant to a gap no other participating project covers.
This is what cross-project standardisation work should look like: each project bringing its unique domain to the table, together building a fuller picture. 🤝
