Becoming Responsible in the Age of AI : Part 1 of 2

A year ago, artificial intelligence was catapulted to the public’s attention with Open AI’s public launch of ChatGPT. For many people, this was the first time they became aware of AI in their daily lives. Across the globe discourse soon swelled with voices championing the transformative value of AI—exponential creativity, new and/or better (but yet to be defined !?!) jobs. Equally loud were the voices of concern—- AI could cause mass unemployment, threaten democracy and “destroy humanity.”?

Ethics, security, privacy, fairness, inclusivity and commitments to avoid “impeding” innovation and competitive advantage became shared themes for both camps. In response many corporations and start-up investors made public commitments to adopt responsible AI practices. Governments across the globe announced or accelerated AI regulatory initiatives. This public and private discourse lay bare the two fundamental challenges–the absence of a shared understanding of “responsible” and the inherent tension (real or imagined) that being “responsible” somehow risked impeding progress.? As we head full speed into the future, here are four key questions all leaders must keep top of mind.

What? exactly does it mean to be responsible??

According to the Oxford Language Dictionary, the word “responsible” is an adjective that means: having an obligation to do something, or having control over or care for someone, as part of one's job or role. It can also mean,? capable of being trusted and/or morally accountable for one’s behavior.? In the? Merriam Webster dictionary, responsible is defined as? liable to be called on to answer, able to answer for one's conduct and obligations; it can also mean being able to choose for oneself between right and wrong. Our lived experiences show us that in practice individuals and societies have different ways of demonstrating obligation, control, trust and morality. Put another way, the term “responsible” has always been subject to interpretation.

Is responsible AI even possible?

“Yes, AND.” The one thing AI detractors and advocates agree on is that there is plenty of room for regulation and no shortage of regulators. A sampling of global regulatory efforts can be found here: United States, European Union, African Union, China. Regulation can provide a shared definition of responsible AI; it is in fact the bare minimum. But becoming responsible in the practice of AI, will lean more on culture and leadership and less on definitions and rules.?

How do we become responsible ?

Certain highly regulated fields, like medicine, have deep experience in operationalizing cultures where ethics and responsibility are core tenets of practice. For example, the Morbidity and Mortality conference or “M&M”, is a weekly conference that occurs in all institutions that have residency programs (physician training programs). “M&M” is mandated by the ACGME (Accreditation Council for Graduate Medical Education). The purpose of the conference is to provide a forum, “ to identify areas of improvement, and promote professionalism, ethical integrity and transparency in assessing and improving patient care. This forum allows a space for teaching curriculum on quality improvement and medico- legal issues to residents and students and to foster a climate of openness and discussion about medical errors” (Source: Emory University School of Medicine).??

Beyond medical practice, in engineering, the Retrospective is a framework that affords space to review and consider work. Unlike “M&M” these are usually self-driven (optional) and often look to assess what was done in the past with the purpose of improving the future. There is no formal review of ethics, medico-legal or related topics (Atalassian).?

In business, the Ethics Canvas ?which leverages the business model canvas template can be used to operationalize ethics practices. The tool enables teams to discuss the ethical implications of their projects, find solutions and act to mitigate harms. A particularly elegant feature of this tool? is that it facilitates a tiered discussion on the impacts of a project facilitating teams to consider impacts at an individual level, a team level, a corporate level up to a societal level. (For example, a product’s negative impact on the company and positive impact on society).

Yet, for many people involved in the building of AI tools, ethics education comes in the form of a once yearly online learning activity (corporate compliance), a few credits in college, self directed learning or none at all. To my knowledge, few workplaces have adopted a systematic, ritual based, practice like “M&M”. In order to become responsible in the Age of AI, organizations will need to operationalize ethics and critique into organizational culture and create incentives that make these practices stick.

What does great leadership look like in the “Age of AI”??

Becoming responsible in the Age of AI will not only require those building and/or working with AI tools to adopt practices similar to M&M or the Ethics canvas. Leaders at all levels of the organization will need to create cultures where the practice of responsible AI is possible. At a minimum hiring, retention, promotion and incentive compensation practices will need to explicitly evaluate leadership in this new dimension. Next week, in Part 2 of this series, Gail Golden, PhD, MBA, will detail the six key qualities leaders will need in the Age of AI. Gail has decades of experience working with businesses to build better leaders as a C-suite executive coach and consultant. Stay tuned!

Read Part 2 of this series, published by Gail Golden, PhD, MBA, here.

Elif E Oker, MD | All Rights Reserved | 2024