In the ever-evolving landscape of recruitment technology, the rise of Artificial Intelligence (AI) has opened up new frontiers in how we approach talent acquisition and job advertising. However, with great power comes great responsibility, especially when it comes to mitigating biases inherent in AI-driven systems. In this context, we're thrilled to bring you an exclusive interview between TAtech Founder & CEO Peter Weddle and Martin Lenz, CEO of Jobiqo.
The recent controversy surrounding the Public Employment Service in Austria (AMS) and its supposedly biased ChatGPT-based chatbot has reignited discussions on the ethical implications of AI in public services. This incident underscores the necessity for trusted institutions to act as intermediaries in the deployment of AI tools, ensuring they are used responsibly and equitably.
In this thought-provoking conversation, Peter and Martin Lenz delve into the potential of AI in reshaping the recruitment landscape, the importance of mitigating biases in AI-driven systems, and Jobiqo's involvement in the "TIMELY" research project aimed at reducing gender bias in job recommender systems. They also discuss the critical role trusted institutions should play in the ethical deployment of AI technologies like ChatGPT, considering insights from AI experts on the challenges of ensuring unbiased AI applications.
Peter Weddle: The recent story about Austria's Public Employment Service and its biased chatbot, built on ChatGPT, has stirred quite a debate. What's your take on the use of Artificial Intelligence (AI) in recruitment advertising and the talent acquisition industry?
Martin Lenz: It's a critical time for AI in our industry. The Austrian case illustrates the potential biases AI can perpetuate if not carefully managed. However, it also highlights AI's capability to transform recruitment by offering personalized, efficient career advice and job matching.
Peter Weddle: How does Jobiqo approach these AI challenges to ensure fairness and effectiveness in job advertising and talent acquisition?
Martin Lenz: At Jobiqo, we prioritize transparency and continuous improvement in our AI models. We actively work to mitigate biases by diversifying our data sources, focusing on non-critical domains to apply AI (e.g. job descriptions), and by applying rigorous testing. We aim to harness AI's power to enhance job search and recruitment advertising processes, making them more inclusive and equitable.
Peter Weddle: I've heard about the "TIMELY" research project that Jobiqo is participating in, which aims to mitigate bias in job recommender systems. Could you tell us more about this initiative?
Martin Lenz: "TIMELY" is a collaborative research project we're proud to be part of, alongside Johannes Kepler University, specifically with the Institute for Computational Perception and the Institute for Legal Gender Studies. The project focuses on identifying and preventing gender bias within job recommender systems. It's a testament to our commitment to tackling the nuanced challenges AI presents, especially concerning bias. By combining our efforts with academic research, we aim to develop more equitable and transparent AI-driven recruitment technologies.
Peter Weddle: That's a proactive approach to a complex issue. Speaking of complexity, you've suggested that trusted institutions like the Austrian Public Employment Service should serve as "gateways" to tools like ChatGPT. Could you elaborate on that?
Martin Lenz: With the growing popularity and interest in AI tools like ChatGPT, it's vital for trusted institutions to mediate these technologies' access. Direct usage by individuals can inadvertently expose them to biases and "hallucinations" inherent in these tools, due to their training on biased datasets and the auto-regressive nature of generating responses based on probabilities. Trusted gateways can implement safeguards, conduct rigorous testing, and ensure ethical use, thus minimizing exposure to these biases.
Peter Weddle: It sounds like a balanced approach to harnessing AI's benefits while mitigating its risks. How do you see this evolving, especially considering AI experts like Meta’s Chief AI Scientist Yann LeCun's views on the inherent biases in Large Language Models (LLMs)?
Martin Lenz: In general, machine learning applications are prone to biases because of the potentially biased data they are trained on. Yann LeCun particularly alerts us on LLM applications (such as ChatGPT) which build on an auto-regressive model architecture and therefore make it very difficult to mitigate all potential bias. This insight shapes our approach to AI development at Jobiqo. We're not just focusing on creating more advanced tools but also on ensuring these tools are transparent, ethical, and as unbiased as possible. It's about continuous improvement and dialogue with the broader community, including regulatory bodies, to navigate these challenges responsibly.
Peter Weddle: Given the criticisms of the Austrian Public Employment Service's chatbot, how should organizations approach developing AI tools for career advice?
Martin Lenz: Organizations should adopt a holistic approach, considering ethical implications and potential biases from the outset. Collaboration with AI ethics experts and clear communication with stakeholders about AI's capabilities and limitations are key. Ultimately, AI should be a tool to augment human decision-making, not replace it. The recent AI Act which will be introduced in the European Union sees recruitment applications as high-risk. So there will be more and more regulations expected around this issue.
Peter Weddle: Lastly, what future developments do you foresee in AI's role within recruitment and job boards?
Martin Lenz: AI will continue to evolve, becoming more sophisticated in understanding complex job requirements and candidate profiles. We'll see AI facilitating not just job matching but also predicting career paths and identifying skills gaps for continuous learning. However, this future hinges on our ability to manage AI responsibly, ensuring it serves to broaden opportunities for all job seekers.
About Martin Lenz
Martin Lenz, CEO of Jobiqo, is renowned for his extensive expertise in the intersection of AI and recruitment technologies, drawing from over two decades of experience in software development, and management consulting, including a significant 9-year tenure at Accenture, and SaaS solutions for job boards. Holding a Master's degree in computer science and additional international business education (Vienna, New York, Hong Kong), he has been responsible for Jobiqo's global growth, making it a leader in job board software and AI-powered recruitment advertising technologies. Beyond his leadership at Jobiqo, Martin is an active contributor to the tech community, holding a board role at the Austrian Computer Society (OCG) and engaging in discussions on labor market transformation. His insights into the ethical deployment of AI in recruitment honed through his career and contributions to organizations like TAtech, position him as a distinguished voice in navigating the challenges and opportunities of AI in talent acquisition.
About Jobiqo
Jobiqo, a leading job board software provider, has been at the forefront of integrating AI into recruitment advertising with an emphasis on ethical use and bias mitigation. With their leading technology, Jobiqo helps job boards, publishers, and associations worldwide, offering an AI-enabled white label job board platform, search and match technology and programmatic job advertising to enable job sites to empower employers to connect with quality candidates. Their commitment to innovation is matched by a dedication to addressing the challenges posed by AI, including potential biases that can arise from its use.