Automated hiring systems are among the fastest-developing of all high-stakes AI systems. Among these are algorithmic personality tests that use insights from psychometric testing, and promise to surface personality traits indicative of future success based on job seekers’ resumes or social media profiles. We interrogate the validity of such systems using stability of the outputs they produce, noting that reliability is a necessary, but not a sufficient, condition for validity. Crucially, rather than challenging or affirming the assumptions made in psychometric testing — that personality is a meaningful and measurable construct, and that personality traits are indicative of future success on the job — we frame our audit methodology around testing the underlying assumptions made by the vendors of the algorithmic personality tests themselves. Our main contribution is the development of a socio-technical framework for auditing the stability of algorithmic systems. This contribution is supplemented with an open-source software library that implements the technical components of the audit, and can be used to conduct similar stability audits of algorithmic systems. We instantiate our framework with the audit of two real-world personality prediction systems, namely, Humantic AI and Crystal. The application of our audit framework demonstrates that both these systems show substantial instability with respect to key facets of measurement, and hence cannot be considered valid testing instruments.
Resume Format, LinkedIn URLs and Other Unexpected Influences on AI Personality Prediction in Hiring: Results of an Audit
Automated hiring systems are among the fastest-developing of all high-stakes AI systems. Among these are algorithmic personality tests that use insights from psychometric testing, and promise to surface personality traits indicative of future success based on job seekers' resumes or social media profiles. We interrogate the reliability of such systems using stability of the outputs they produce, noting that reliability is a necessary, but not a sufficient, condition for validity. We develop a methodology for an external audit of stability of algorithmic personality tests, and instantiate this methodology in an audit of two systems, Humantic AI and Crystal. Rather than challenging or affirming the assumptions made in psychometric testing --- that personality traits are meaningful and measurable constructs, and that they are indicative of future success on the job --- we frame our methodology around testing the underlying assumptions made by the vendors of the algorithmic personality tests themselves.
In our audit of Humantic AI and Crystal, we find that both systems show substantial instability on key facets of measurement, and so cannot be considered valid testing instruments. For example, Crystal frequently computes different personality scores if the same resume is given in PDF vs. in raw text, violating the assumption that the output of an algorithmic personality test is stable across job-irrelevant input variations. Among other notable findings is evidence of persistent --- and often incorrect --- data linkage by Humantic AI.
An open-source implementation of our auditing methodology, and of the audits of Humantic AI and Crystal, is available at https://github.com/DataResponsibly/hiring-stability-audit.
more »
« less
- NSF-PAR ID:
- 10352861
- Date Published:
- Journal Name:
- AIES '22: Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society
- Page Range / eLocation ID:
- 572 to 587
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
more » « less
Abstract -
null (Ed.)The high bar of proof to demonstrate either a disparate treatment or disparate impact cause of action under Title VII of the Civil Rights Act, coupled with the “black box” nature of many automated hiring systems, renders the detection and redress of bias in such algorithmic systems difficult. This Article, with contributions at the intersection of administrative law, employment & labor law, and law & technology, makes the central claim that the automation of hiring both facilitates and obfuscates employment discrimination. That phenomenon and the deployment of intellectual property law as a shield against the scrutiny of automated systems combine to form an insurmountable obstacle for disparate impact claimants. To ensure against the identified “bias in, bias out” phenomenon associated with automated decision-making, I argue that the employer’s affirmative duty of care as posited by other legal scholars creates “an auditing imperative” for algorithmic hiring systems. This auditing imperative mandates both internal and external audits of automated hiring systems, as well as record-keeping initiatives for job applications. Such audit requirements have precedent in other areas of law, as they are not dissimilar to the Occupational Safety and Health Administration (OSHA) audits in labor law or the Sarbanes-Oxley Act audit requirements in securities law. I also propose that employers that have subjected their automated hiring platforms to external audits could receive a certification mark, “the Fair Automated Hiring Mark,” which would serve to positively distinguish them in the labor market. Labor law mechanisms such as collective bargaining could be an effective approach to combating the bias in automated hiring by establishing criteria for the data deployed in automated employment decision-making and creating standards for the protection and portability of said data. The Article concludes by noting that automated hiring, which captures a vast array of applicant data, merits greater legal oversight given the potential for “algorithmic blackballing,” a phenomenon that could continue to thwart many applicants’ future job bids.more » « less
-
null (Ed.)A received wisdom is that automated decision-making serves as an anti-bias intervention. The conceit is that removing humans from the decision-making process will also eliminate human bias. The paradox, however, is that in some instances, automated decision-making has served to replicate and amplify bias. With a case study of the algorithmic capture of hiring as heuristic device, this Article provides a taxonomy of problematic features associated with algorithmic decision-making as anti-bias intervention and argues that those features are at odds with the fundamental principle of equal opportunity in employment. To examine these problematic features within the context of algorithmic hiring and to explore potential legal approaches to rectifying them, the Article brings together two streams of legal scholarship: law and technology studies and employment & labor law. Counterintuitively, the Article contends that the framing of algorithmic bias as a technical problem is misguided. Rather, the Article’s central claim is that bias is introduced in the hiring process, in large part, due to an American legal tradition of deference to employers, especially allowing for such nebulous hiring criterion as “cultural fit.” The Article observes the lack of legal frameworks that take into account the emerging technological capabilities of hiring tools which make it difficult to detect disparate impact. The Article thus argues for a re-thinking of legal frameworks that take into account both the liability of employers and those of the makers of algorithmic hiring systems who, as brokers, owe a fiduciary duty of care. Particularly related to Title VII, the Article proposes that in legal reasoning corollary to extant tort doctrines, an employer’s failure to audit and correct its automated hiring platforms for disparate impact could serve as prima facie evidence of discriminatory intent, leading to the development of the doctrine of discrimination per se. The article also considers other approaches separate from employment law such as establishing consumer legal protections for job applicants that would mandate their access to the dossier of information consulted by automated hiring systems in making the employment decision.more » « less
-
null (Ed.)Purpose The purpose of this paper is to offer a critical analysis of talent acquisition software and its potential for fostering equity in the hiring process for underrepresented IT professionals. The under-representation of women, African-American and Latinx professionals in the IT workforce is a longstanding issue that contributes to and is impacted by algorithmic bias. Design/methodology/approach Sources of algorithmic bias in talent acquisition software are presented. Feminist design thinking is presented as a theoretical lens for mitigating algorithmic bias. Findings Data are just one tool for recruiters to use; human expertise is still necessary. Even well-intentioned algorithms are not neutral and should be audited for morally and legally unacceptable decisions. Feminist design thinking provides a theoretical framework for considering equity in the hiring decisions made by talent acquisition systems and their users. Social implications This research implies that algorithms may serve to codify deep-seated biases, making IT work environments just as homogeneous as they are currently. If bias exists in talent acquisition software, the potential for propagating inequity and harm is far more significant and widespread due to the homogeneity of the specialists creating artificial intelligence (AI) systems. Originality/value This work uses equity as a central concept for considering algorithmic bias in talent acquisition. Feminist design thinking provides a framework for fostering a richer understanding of what fairness means and evaluating how AI software might impact marginalized populations.more » « less
-
Drawing, as a skill, is closely tied to many creative fields and it is a unique practice for every individual. Drawing has been shown to improve cognitive and communicative abilities, such as visual communication, problem-solving skills, students’ academic achievement, awareness of and attention to surrounding details, and sharpened analytical skills. Drawing also stimulates both sides of the brain and improves peripheral skills of writing, 3-D spatial recognition, critical thinking, and brainstorming. People are often exposed to drawing as children, drawing their families, their houses, animals, and, most notably, their imaginative ideas. These skills develop over time naturally to some extent, however, while the base concept of drawing is a basic skill, the mastery of this skill requires extensive practice and it can often be significantly impacted by the self-efficacy of an individual. Sketchtivity is an AI tool developed by Texas A&M University to facilitate the growth of drawing skills and track their performance. Sketching skill development depends in part on students’ self-efficacy associated with their drawing abilities. Gauging the drawing self-efficacy of individuals is critical in understanding the impact that this drawing practice has had with this new novel instrument, especially in contrast to traditional practicing methods. It may also be very useful for other researchers, educators, and technologists. This study reports the development and initial validation of a new 13-item measure that assesses perceived drawing self efficacy. The13 items to measure drawing self efficacy were developed based on Bandura’s guide for constructing Self-Efficacy Scales. The participants in the study consisted of 222 high school students from engineering, art, and pre-calculus classes. Internal consistency of the 13 observed items were found to be very high (Cronbach alpha: 0.943), indicating a high reliability of the scale. Exploratory Factor Analysis was performed to further investigate the variance among the 13 observed items, to find the underlying latent factors that influenced the observed items, and to see if the items needed revision. We found that a three model was the best fit for our data, given fit statistics and model interpretability. The factors are: Factor 1: Self-efficacy with respect to drawing specific objects; Factor 2: Self-efficacy with respect to drawing practically to solve problems, communicating with others, and brainstorming ideas; Factor 3: Self-efficacy with respect to drawing to create, express ideas, and use one’s imagination. An alternative four-factor model is also discussed. The purpose of our study is to inform interventions that increase self-efficacy. We believe that this assessment will be valuable especially for education researchers who implement AI-based tools to measure drawing skills.This initial validity study shows promising results for a new measure of drawing self-efficacy. Further validation with new populations and drawing classes is needed to support its use, and further psychometric testing of item-level performance. In the future, this self-efficacy assessment could be used by teachers and researchers to guide instructional interventions meant to increase drawing self-efficacy.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3514094.3534189
