Advancing AI ethics beyond compliance

If you learned that a customer’s credit application had been rejected with no discernible justification, would you condone the decision? If a doctor with no access to the latest medical literature recommended that a loved one undergo an invasive procedure, would you authorize the surgery? Surely not.

Critical areas of judgment – especially decisions that directly impact others’ lives and well-being – are governed by standards of appropriate action. Humans live by communally governed ethical norms, enforced by laws, rules, societal pressures, and public discourse. While ethics may vary over time and across cultures, they have played a crucial role in guiding decisions since early human civilization.

Video: Who is responsible for integrating ethics into AI?

Envisioning AI ethics

In business, the topic of ethics is also not new. Traditionally, ethics are one of the much-talked-about but often-ignored elements of modern business culture. In the quest for financial success, sometimes corners are cut, advantages are taken, and long-term priorities – and even values – are sacrificed for short-term gains. In response, a compliance apparatus has been created inside companies. Organizations have worked to create “guardrails” and other reinforcement mechanisms to combat lapses, whether inadvertent or willful.

Arguably, ethical considerations have never been more critical than they are today. People around the world, from business executives and front-line employees to government ministers and individual citizens, have found themselves confronting critical decisions they wouldn’t have imagined making in the past – decisions that could profoundly impact the lives of their fellow colleagues and citizens. And many have been forced to weigh seemingly impossible trade-offs between economic and health imperatives, guided only by their ethics, morals, and values.

Yet the existing system is ill-equipped for the challenges to come. In the last few years, the business environment has been rapidly changing, with a growing number of decisions facilitated by artificial intelligence (AI). Companies now use AI to help conduct talent screening, insurance claims processing, customer service, and a host of other important workflows. But the ethical parameters around AI remain vague and intangible, in some instances pushed aside as impediments to progress.

AI ethics: The corporate landscape

Meanwhile, AI adoption is rapidly growing. And with heightened AI use comes heightened risk, in areas ranging from data responsibility to inclusion and algorithmic accountability.

To gain a deeper understanding of executive views on the consequences and ethical considerations associated with AI’s growing role in the enterprise, we sought input from 1,250 executives from a variety of industries and geographies. Our research suggests AI’s importance to organizational strategy is likely to double in the next three years, underscoring the need to address the topic of ethics.

The unavoidable conclusion: Ethical considerations must be elevated in the dialogue about AI systems across the business landscape. The level of cognitive understanding between humans and machines is inherently lower than it is between humans and other humans, yet the latter arena has been structured for centuries around ethics. Since AI relies on huge computing power, it can derive insight from massive amounts of data that would challenge human cognition. Relying only on traditional ethical approaches to decision making may be insufficient in addressing AI-powered decisions.

Fostering a sustainable future

More than half of the executives we surveyed tell us AI actually could improve their companies’ ethical decisions. A majority also say AI could be harnessed as a force for societal good, not just for good business. When one chief human resource officer was asked what came to mind when thinking about AI, the answer was “advancement in technology for betterment of human life,” a sentiment echoed by others.

While xenophobia, misogyny, and other biases – even if unintentional – can be obscured behind human rationales, study respondents suggest that AI can be designed with fairness, transparency, and even empathy. Detecting and correcting bias in AI – teaching technology to be more effective in relating to humans – may advance organizations’ abilities to work together and achieve greater outcomes.

AI offers the opportunity to diagnose root causes of unintended results, essentially debugging biases. This could enable better understanding of past shortcomings and improvement in achieving social goals. But first, the right ethical frameworks have to be in place.

Read the full report to learn more about executives’ thoughts on AI ethics and how organizations can proactively address the topic while shaping their competitive future.