DEI in AI (ChatGPT & Google Bard)

A practical comparison of ChatGPT and Google Gemini (formerly Bard) looks at how each platform can be applied to sourcing and diversity recruiting work, with specific examples pulled from real searches rather than general claims about artificial intelligence.

One demonstration walks through converting a military resume into civilian language using ChatGPT. A resume listing a "soldier" job title and bullet points about weapon systems and equipment accountability gets rewritten into terms like "operations specialist" and "expertise in equipment and maintenance systems," making it easier for recruiters unfamiliar with military rank structures and job codes to identify transferable skills.

Boolean string generation is tested on both platforms. ChatGPT handles site-command searches well, producing strings that can be pasted directly into Google to surface candidate profiles tied to organizations such as the National Society of Black Engineers. Gemini struggles with site commands but compensates with stronger plugin integration, connecting to Sheets and other Google tools, and returning search results with direct hyperlinks to source pages.

Other examples include:

• Generating a full list of historically Black colleges and universities and converting it into a ready-to-use Boolean string for LinkedIn Recruiter, cutting out manual research time
• Building layered Boolean strings for veteran sourcing, starting with basic rank and title terms and expanding into a fuller list of military-specific language
• Filtering broad keyword search results, which can return millions of profiles, down by location, job title, or current employer to reach a workable candidate list
• Comparing subscription costs and feature sets, including free versus paid tiers and the tradeoffs between ChatGPT's plugin ecosystem and Gemini's

The larger point running through each example is that these tools work best as a starting point rather than a finished answer. A generated Boolean string or a translated resume still needs a recruiter to filter, verify, and apply judgment before reaching out to candidates. Used that way, either platform can meaningfully speed up the work of finding and engaging veterans, HBCU graduates, and other underrepresented candidates, while also helping recruiters catch language in job descriptions or search strings that may be excluding the people they are trying to reach.