In many traditional selection processes, if you don't have a resume with "status," experience at recognized companies, or strong networks, you're already behind—even before you're seen. This reproduces privileges and excludes talent that could have the most authentic impact. But technology is opening doors. Well-applied AI and the use of structured data can change this dynamic. Here we explore how, why it works, and what precautions to take.

What studies show about biases embedded in recruitment

- A classic study is that of the NBER, Employers' Replies to Racial Names (Bertrand & Mullainathan, 2003), which showed that candidates with names with African-American connotations need to send more resumes to receive the same number of invitations as candidates with ostensibly "white" names. NBER

- Another study by MIT Sloan (Danielle Li et al., “Hiring as Exploration”) shows that well-designed algorithms (which include exploration, i.e., do not just reproduce history) increase both the quality and diversity of selected candidates. MIT Sloan

- There is also recent research that identifies how AI models used in resume screening or automated interviews continue to reflect biases—based on name, ethnicity, gender, etc. For example, tools that evaluate names associated with different races tend to favor "white" names in many contexts.

These biases are not merely unfair: they have a real cost, both for candidates (who become demotivated, disengaged, and excluded) and for companies (loss of talent, reputation, and diversity of thought).

How AI + data can level this playing field

Here are some practical points on how Solu and similar solutions have been operating:

Fair evaluation of everyone: AI makes it possible to interview all applicants and evaluate them beyond their resumes—what they have learned outside of traditional companies, participation in personal or community projects, reasoning and communication skills, etc.

Reduction of superficial filters: Instead of relying solely on keywords or "having worked at X company," behavior, clarity of thought, motivation, and expectations are evaluated. This provides opportunities for those who have had unconventional careers, reinvented themselves, or taken courses outside the formal system.

Actionable data and transparency: With recordings/summaries/reports, the company has visibility into what really motivated the decision (who passed, who didn't, why). This allows for better conversations with candidates, feedback, and adjustments to practices.

Increased diversity and inclusion: When candidates who were normally excluded—based on their name, educational background, or CV—are included based on more humane criteria, it is common to see an improvement in diversity. Not only is this fair, but it also brings different perspectives, greater innovation, and engagement.

Essential precautions for this to work well

To prevent AI from becoming a new exclusionary filter or repeating injustices, it is essential to:

- Train the model with diverse data, with varied contexts, avoiding overfitting in "traditional" profiles.

- Monitor bias: conduct regular audits to see if certain groups (by name, gender, ethnicity, etc.) are consistently being disadvantaged.

- Transparency with candidates: explain how AI works, what criteria are used, and how they will be evaluated.

- Real feedback: not just "approved/not approved," but insights into what worked and what could be improved.

- Involve humans in final decisions, especially to validate cultural fit, values, and expectations.

Practical examples of application

- A Solu client who was already using ATS realized that their shortlists were predictable and lacked diversity. By integrating Solu's screening process, they discovered candidates with non-traditional profiles who were highly aligned with what mattered for the position.

- In internship or trainee processes, where many candidates do not have robust resumes, AI interviews can be used to learn about motivations, ideas, and expectations—elements that show potential where the CV does not.

AI + data are not a silver bullet, but they have the real capacity to level the playing field. They give voice, reduce injustice, and reveal invisible talents. Companies that adopt this approach not only "solve an ethical problem," but also gain a strategic advantage: more diverse teams, more informed decisions, a positive reputation, and fewer mistakes in the hiring process. While old selection processes favored those who already had visibility, this new approach honestly favors those who have potential.

‍