Gen AI: Why Organizations Need Bias-Free SystemsEnrico Panai on Imperatives of a Multi-Disciplinary Approach in Business AI Systems
In the rapidly evolving digital landscape, the integrity of AI systems can make or break an enterprise's technological progress. With the advent of generative AI, it is even more crucial to ensure its ethical use for sustainable progress.
Amid the dynamic evolution of ethics in response to technological advancements, organizations face the challenge of keeping pace with the rapidly advancing generative AI, which presents an extended scope of application and an unprecedented level of complexity.
Enrico Panai, professor of AI ethics at Emlyon Business School and La Cattolica, as well as the president of the Association of AI Ethicists, argues that the challenge lies not in the speed of generative AI's evolution, but in establishing robust safety protocols and setting clear, transparent objectives.
Much like any other technology, proper application and the involvement of qualified experts are crucial. Drawing a pertinent analogy, Panai underscored that just as one wouldn't solely rely on engineers to develop AI for children without input from child psychologists, a multi-disciplinary approach is imperative for business AI systems.
"If those developing technology aim for ethics and do not use the right tools, they risk embedding their beliefs and individual biases into the technology. Organizations need a collective and scientific approach" Panai said. "Even if it's not hard science like maths or physics, we do have tools that can help make a good compromise. Hence, the need for ethicists in every development team."
In this interview with ISMG, Panai also discussed:
- AI ethics and generative AI;
- Why organizations need trained AI ethicists;
- The impact of the EU AI Act on ethics.
Panai is an AI ethicist and human information interaction specialist. He teaches AI ethics at leading universities in France and Italy. He is a member of the French standardization committee for IA – AFNOR and co-convenor on AI-enhanced nudging at CEN-CENELEC.