this section will have annotations for refrenced articles.
| “AI is reinventing hiring—with the same old biases” | |||
|---|---|---|---|
| What it covers | This article explores how AI systems act out human biases instead of fixing the problem of bias. This article provides examples involving race, gender, education and employment gaps and explains how biased historical data can influence the data we are feeding the AI. it ends up creating decisions based on bias and enforces it throughout the hiring practices. | ||
| Why it's valuable | The article challenges the assumption that ai is automatically objective and the importance of checking all the data for AI hiring practices. | ||
| Who would benefit | The people that would benefit from this are prospective companies, researchers studying AI and humans, HR and job seekers. | ||
| limitations | The limitation of the article is that its just an article written with opinion and does not have full on academic research. | ||
| Citation | Here’s how to avoid that trap. MIT Sloan School of Management. https://mitsloan.mit.edu/ideas-made-to-matter/ai-reinventing-hiring-same-old-biases-heres-how-to-avoid-trap | ||
| “When algorithms decide who belongs” | |||
|---|---|---|---|
| What it covers | OpenGlobalRights writes about how AI recruitment systems can make disadvantages for people with disabilities when the system recognizes educational gaps, unusual career paths or differences in communication. This flags the system and basically sees them as negative indicators for a candidate. | ||
| Why it's valuable | The information here is valuable because it expands the discussion of AI bias farther than just race and gender. Very many articles focus on that and not many speak of the bias towards a different group. | ||
| Who would benefit | This benefits human rights researchers or technology research, recruiters and policymakers. | ||
| limitations | The limitations of the article is that it does only focuses on the bias towards one specific group of people. | ||
| Citation | Gupta, V. (2026, September 8). When algorithms decide who belongs: Can AI undermine disability rights?OpenGlobalRights. https://www.openglobalrights.org/when-algorithms-decide-who-belongs-can-ai-undermine-disability-rights/ | ||
| “Amazon Scraps Secret AI Recruiting Tool That Showed Bias Against Women” | |||
|---|---|---|---|
| What it covers | This article investigates Amazon's AI recruiting experiment that ended up being abandoned due to its bias. The system learned patterns from historical resumes and reportedly chose languages associated with male engineers. The system rejected any resume with references to women and created unreliable recommendations. This led to Amazon shutting down the project. | ||
| Why it's valuable | This article is important because it shows a real life example of how AI can affect hiring practices. It's a concrete corporate case study proving how using historical data can produce bias. We as humans are biased but this proves that we must fix the data we feed the AI systems and how connected we are to those biases in the systems. | ||
| Who would benefit | This article is useful for companies wanting to utilize AI, researchers, possible candidates and students preparing for jobs after college. | ||
| limitations | The limitations this article has is that it only focuses on one real life situation of AI bias. It's also a bit outdated as it was reported on the 2018 systems. It can not be treated as a representation of the whole or current AI hiring technologies. | ||
| Citation | Dastin, J. (2018, October 11). Insight—Amazon scraps secret AI recruiting tool that showed bias against women. Reuters. https://www.reuters.com/article/world/insight-amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK0AG/ | ||
| “ 90% of Companies Use AI in Hiring. Fewer Than 5% Are Seeing It Work” | |||
|---|---|---|---|
| What it covers | This article examines research from ManpowerGroup Talent Solutions and developed by Everst group on AI systems in Hiring. It mentions the major gap between widespread integration and transformation in hiring. More than 90% of the surveyed groups use AI while there's fewer than 5% that say the outcome is “transformational”. | ||
| Why it's valuable | This article is valuable because it helps us understand the amount of companies who have adopted AI hiring systems as well as the generated candidate behavior and workflow integration. | ||
| Who would benefit | this benefits HR, researchers and prospective companies. | ||
| limitations | the limitations are that the survey is kind of small as it surveys only 80 senior leaders and it relies on self reporting. | ||
| Citation | Yahoo Finance. (2026, June 23). 90% of companies use AI in hiring. Fewer than 5% are seeing it work. https://finance.yahoo.com/technology/ai/articles/90-companies-ai-hiring-fewer-165900794.html | ||
| “AI Hiring Tools Don’t Just Learn Bias, AI Forms New Biases of Its Own ” | |||
|---|---|---|---|
| What it covers | The article covers a more recent study (2026) suggesting that AI can develop new and factually unsupportive biases even when being trained off of natural information. It was found that in AI models like ChatGPT, Claude and Gemini can develop stereotypes about different fictional demographics during repeated hiring tasks. | ||
| Why it's valuable | The article is valuable because it expands on how even if we use neutral data the AI can become biased from system usage. | ||
| Who would benefit | It benefits AI researchers, HR, ethics researchers and candidates. | ||
| limitations | The limitations of the article is that it uses fictional groups as well as a controlled environment and doesn't work on real hiring environments. | ||
| Citation | Travis, M. (2026, August 11). AI hiring tools don’t just learn bias, AI forms new biases of its own. Forbes. https://www.forbes.com/sites/michelletravis/2026/08/11/ai-hiring-tools-dont-just-learn-bias-ai-forms-new-biases-of-its-own/ | ||
| “AI Hiring Tools May Favor Their Own Work, Smith Study Finds ” | |||
|---|---|---|---|
| What it covers | The article summarizes research investigating whether AI hiring favors certain resumes made by another AI system that's used to evaluate candidates. It studies more than 2200 resumes and reports that the systems favor their own AI material more than any human made or competing models. | ||
| Why it's valuable | The value of the article is that it writes about a less talked about form of hiring bias and also shows how different AI systems interact with each other. | ||
| Who would benefit | This article benefits tech researchers, employers, HR and candidates. | ||
| limitations | The limitation of the article is that the research comes from a controlled environment rather than a broad observational data. | ||
| Citation | Robert H. Smith School of Business. (2026). AI hiring tools may favor their own work, Smith study finds. University of Maryland. https://www.rhsmith.umd.edu/news/ai-hiring-tools-may-favor-their-own-work-smith-study-finds | ||
| “Employer Branding News — AI in Hiring Statistics 2026” | |||
|---|---|---|---|
| What it covers | The article compiles statistics from 2026 on AI systems being adopted by companies to help them in hiring. There's also data on algorithmic bias, if applicants actually trust the systems, regulation and policies on AI and how many applicants are actually applying to sites using AI hiring practices. | ||
| Why it's valuable | The information on this article is important as it brings different forms of data from multiple groups to give an overview on AI Hiring. | ||
| Who would benefit | This benefits HR, researchers and companies who want more data on AI and its popularity. | ||
| limitations | The data is limited because it gets all of its information from a collection of sources rather than its own research. | ||
| Citation | Robbins, J. (2026, June 9). AI in hiring statistics 2026: Adoption, bias, trust, and regulation. Employer Branding News. https://employerbranding.news/resources/ai-in-hiring-statistics-2026-adoption-bias-trust-and-regulation/ | ||
| “ Unilever Saves on Recruiters by Using AI to Assess Job Interviews” | |||
|---|---|---|---|
| What it covers | The Guardian article focuses on Unilever’s use of HireVue software to evaluate graduate candidates through video interviews. Language, tone, facial expressions, and body language is viewed and studied by the system. It reports efficiency gains, like reductions in recruitment time and cost. It also presents concerts about transparency, public acceptance, and potential discrimination in the automated systems decision making. | ||
| Why it's valuable | This article is valuable because it shows an early and concrete example of AI being integrated into recruitment rather than just hypotheticals and theoretical risks. | ||
| Who would benefit | The information could be utilized by companies considering using AI, recruiters, researchers and possible graduate candidates. | ||
| limitations | The limitations of the article is that there's not much evidence of effectiveness on the Hirevue system. It claims to promote diversity but doesn't have the actual data to prove this. Since the AI evaluates body language, tone and expressions there is potential bias but the article does not go into detail about the bias of the system. The article is also a bit outdated since it was written in 2019 and since these systems have advanced a lot, it doesn't reflect the current systems. | ||
| Citation | Booth, R. (2019, October 25). Unilever saves on recruiters by using AI to assess job interviews. The Guardian. https://www.theguardian.com/technology/2019/oct/25/unilever-saves-on-recruiters-by-using-ai-to-assess-job-interviews | ||