In a recent experiment sparked by an intriguing online thread, the question was raised whether artificial intelligence (AI) perpetuates gender bias when drafting resumes, particularly within the legal profession. Jen Horsburgh's viral tweet revealed how Google’s Gemini AI rephrased her resume differently under a male name, hinting at underlying biases in AI systems trained on historically skewed data.

The legal industry, where resumes play a pivotal role in career advancement, became the focus of a further test. In our investigation, identical profiles of a fictitious fifth-year M&A associate named "Vivian" and "Thomas" were fed into Gemini, keeping all variables constant except for the name. The output was telling.

While the AI managed to avoid blatant discrepancies in job descriptions—both versions included terms like "represented," "advised," and "managed"—subtler biases emerged in the professional summaries and other sections. Thomas was described with added flair as a "High-performing" associate with a "Proven track record," whereas Vivian's summary was straightforward and less embellished.

Moreover, Thomas's resume highlighted his pro bono work prominently, while entirely omitting Vivian's similar efforts. Educational achievements and extracurricular activities were also presented differently; Thomas's credentials included detailed accolades, whereas Vivian's were less emphasized.

This experiment underscores the subtle yet impactful ways AI can perpetuate gender biases, even when handling identical data. While legal terminology's standardization might shield against the most overt biases, the nuances of AI-generated text still require scrutiny.

Legal professionals should be wary of relying solely on AI for resume creation. Variations in the presentation of identical qualifications can influence hiring decisions, potentially affecting career trajectories. Running multiple drafts and actively reviewing AI outputs is advisable to ensure fair and accurate representation.

As AI continues to integrate into professional settings, ongoing vigilance and adjustments will be crucial in addressing and mitigating embedded biases. The legal industry, with its reliance on precise language and documented achievements, is particularly positioned to lead by example in refining AI applications for a fairer future.