HR’s Golden Moment: How HR Can Lead Their Organizations in Defining AI Governance
[1:54] Why is it HR’s Golden Moment?
Keith Sonderling, EEOC Commissioner, and Athena Karp, Founder and CEO of HiredScore, underscore the current uncertainty within companies, especially large global organizations, concerning the governance of Artificial Intelligence technologies. AI is increasingly being integrated into various departments and divisions, with a notable percentage of executive teams earmarking funds for AI programs and expansions throughout their organizations.
This period is particularly noteworthy due to the unique opportunity for HR to potentially take the lead in addressing one of the most complex and unresolved questions at the CTO, CIO, CEO, and Board levels – namely, how to identify instances where AI can be effectively employed and how to establish robust governance frameworks for those specific use cases. HR is uniquely positioned to emerge as a beacon within the organization in tackling this crucial aspect.
Why is that so? The evolution of the HR function in the last couple of years has been substantial, especially in response to significant societal events such as the Me Too movement, George Floyd's case, and the COVID-19 pandemic. These events have propelled HR departments to the forefront, enhancing their roles within organizations. Keith at the EEOC, views this development positively, as it shifts issues traditionally considered administrative to the forefront of business attention.
For instance, the Me Too movement prompted HR to proactively address issues of sexual harassment, with high-profile dismissals underscoring the importance of enforcing policies regardless of an individual's status. With the onset of COVID-19, the focus shifted to employee safety and adapting to a rapidly changing workforce landscape. Major consultancy firms have also acknowledged HR's pivotal role in guiding organizations through these challenges, impacting both employee well-being and business performance.
Keith Sonderling talks about the growing integration of AI in HR processes. AI is becoming pervasive in every aspect of the employment relationship, from crafting job descriptions to handling terminations. Keith Sonderling emphasizes that using AI in HR significantly differs from other business applications, as it involves managing people's livelihoods and civil rights. By adopting AI responsibly and creating governance frameworks for the use of AI in HR, HR teams have a “golden opportunity” to take the lead in AI governance within organizations, showcasing their value and expertise beyond traditional HR functions.
[9:13] AI Governance in HR - Are the regulations actually that different for AI in HR?
The EEOC is focused on ensuring that all employment decisions, especially those influenced by technology, comply with legal standards. Keith mentions that while the people at the EEOC are not technologists, their understanding of employment decisions aligns closely with that of HR professionals. Their primary concern is whether employment decisions are based on the requisite skills and qualifications for a job or if they are tainted by discrimination, whether intentional or inadvertent through unnecessary job requirements that disproportionately exclude certain protected classes.
As HR has always done, we must remember that technological tools, regardless of their sophistication, are ultimately aids in making employment decisions. These tools, which encompass AI, machine learning, and other advanced algorithms, can range from being assistive to making decisions autonomously. However, the essence of HR decision-making remains unchanged from the pre-1960s era when EEOC was established. The novelty lies not in the nature of the decisions but in the methods and transparency with which they are now made.
The challenge we face is understanding these technologies within the context of HR's traditional decision-making framework. Organizations need to recognize that the integration of algorithmic technologies into HR processes—be it in hiring, promotions, salary determinations, or performance evaluations—is not creating new types of decisions but augmenting or automating existing ones. This integration should be guided by the same legal and ethical principles that have always governed HR practices.
However, the advent of technology in HR brings both opportunities and complexities. The abundance of data and the "black box" nature of some algorithms can be daunting. It is vital to prioritize understanding how AI tools impact the range of employment decisions within an organization. The decisions facilitated by these tools are subject to the same legal scrutiny as those made without technological assistance. This continuity is crucial in EEOC’s regulatory approach.
Moreover, technology can enhance transparency and fairness in employment decisions. In cases of alleged discrimination, technology can provide a more tangible and auditable trail of decision-making processes than traditional methods. It allows organizations to demonstrate compliance with legal and ethical standards more effectively.
However, leveraging technology in HR is not without its challenges. It requires a significant upfront investment in understanding these tools appropriately. It's not a "set it and forget it" scenario; it involves ongoing engagement and refinement based on upcoming AI laws. In this evolving landscape, any regulatory body, including EEOC’s aim is not to stifle innovation but to ensure that technological advancements in HR are aligned with legal compliance and ethical governance. This approach enables organizations to harness the potential of technology to make well-informed, transparent, and fair employment decisions.
[23:00] Challenges and responsibilities HR departments face in ensuring non-discriminatory practices, particularly in the context of using AI and algorithms for hiring
- Identifying Discrimination: It's essential to recognize whether employment systems are discriminatory, intentionally or not. This involves examining the diversity of data sets used, such as applicant pools, current employee demographics, and the broader labor market.
- Obligations of Federal Contractors: For federal contractors, there's a specific affirmative obligation to ensure non-discriminatory practices, as mandated by the OFCCP (Office of Federal Contract Compliance Programs).
- Input Data and Algorithmic Bias: Care must be taken in selecting the data and criteria fed into algorithms. Even with diverse data, biases can arise if incorrect skills or qualifications are emphasized.
- Job Qualifications and Standards: Long-standing industry standards and job descriptions may inherently carry biases. Each requirement should be scrutinized to ensure it's necessary and doesn't unintentionally exclude certain groups.
- Updating Skills and Requirements: Technology allows for rapid updating of job skills and requirements. Adjusting qualifications, like reducing experience requirements and offering internal training, can broaden the applicant pool and reduce discrimination.
- Control Over Algorithms: It's crucial to monitor who controls the algorithm and the weight given to various factors. There's a risk of intentional discrimination if biases are introduced at this stage.
- Proactive Measures by HR Departments: HR departments should proactively work to prevent discrimination at every step, from data selection to algorithmic design and job qualification setting. This involves continuous effort and self-regulation, without waiting for external accusations or new legislation.
- Benefits of Diligence: Thorough and conscientious practices can lead to less discrimination and greater diversity in hiring, giving opportunities to individuals who might otherwise be overlooked due to systemic biases.
For additional insights on strategic approaches to AI in HR, including considerations for choosing between comprehensive AI solutions versus specialized, use-case-specific tools, and key questions to ask AI vendors, be sure to view the complete webinar embedded above.