{"id":253,"date":"2026-03-18T10:43:55","date_gmt":"2026-03-18T10:43:55","guid":{"rendered":"https:\/\/ai-rebels.uhu.nl\/?p=253"},"modified":"2026-03-26T15:41:23","modified_gmt":"2026-03-26T15:41:23","slug":"ai-and-diversity-inclusion","status":"publish","type":"post","link":"https:\/\/ai-rebels.uhu.nl\/en\/ai-en-diversiteit-inclusie\/","title":{"rendered":"AI and diversity & inclusion"},"content":{"rendered":"
<\/div><\/div><\/div>
\n\t
\n\t\t
\n\t\t\t
\n\t\t\t\t\n\t\t\t<\/div> \n\t\t<\/div>\n\t<\/div> \n\n\t
\n\t\t
\n\t\t\t
\n\t\t\t\t\n
\n\t

How do you make sure your D&I story is correctly represented by LLMs?<\/h2>\n

The representation dilemma<\/h3>\n

A candidate asks Claude: \u201cWhich tech companies in the Netherlands are good for female developers?\u201d<\/p>\n

Claude searches the web. It finds your D&I page (\u201cWe believe in diversity\u201d). It finds a news item about your female CTO. It finds a Glassdoor review that says \u201cmen's stronghold.\u201d It finds statistics you never published.<\/p>\n

What will be the answer? That depends on which signals are strongest, and you can influence that.<\/p>\n

AI and bias: the double challenge<\/h3>\n

The challenge is twofold:<\/p>\n

1. AI systems can contain bias<\/strong><\/p>\n

LLMs are trained on historical data that reflects societal biases. Research shows that AI systems in recruitment\u00a0gender and racial bias may be present<\/strong> in CV screening, patterns that may also carry over into how they describe employers.<\/p>\n

2. Your D&I story must go through the AI filter<\/strong><\/p>\n

Even if your D&I efforts are excellent, that information needs to be findable, citable and compelling for AI systems.<\/p>\n

What works: data over claims<\/h3>\n

AI systems (like critical candidates) are sceptical of vague D&I claims. What does work:<\/p>\n\n\n\n\n\n\n\n
Weak (unquote)<\/th>\nStrong (quotable)<\/th>\n<\/tr>\n
\u201cWe believe in diversity\u201d<\/td>\n\u201c42% of our tech teams are women (2025)\u201d<\/td>\n<\/tr>\n
\u201cInclusive workplace\u201d<\/td>\n\u201cEmployee Resource Groups for five communities\u201d<\/td>\n<\/tr>\n
\u201cEqual opportunities\u201d<\/td>\n\u201cPay gap audit: 2.3% difference, action plan published\u201d<\/td>\n<\/tr>\n
\u201cDiverse leadership\u201d<\/td>\n\u201c4 of 8 board members are women; 2 have migration background\u201d<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The rule:\u00a0specific, verifiable data is cited; vague claims are ignored<\/strong>.<\/p>\n

AI tools for D&I in recruitment<\/h3>\n

Interesting: AI itself is being used to reduce bias. Research shows effective applications:<\/p>\n