Foreign Corrupt Practices Act

Ethical conduct is about more than risk management processes and regulatory box-ticking. When a company establishes clear values and ethical guidelines and systematically reinforces them at every level, it communicates to stakeholders broadly what the organization stands for and serves as a North Star for employees as they navigate decisions, protecting one of a company’s core assets: its reputation. Ethics-driven behavior is the invisible connection between a company’s fundamental values and its day-to-day operations, driving not only sound, consistent decisions but also ultimately improved performance.

Often, however, companies face challenges that are not black and white, that occur in the gray areas where corporate priorities collide, or norms or regulations are still uncertain or emerging. This is particularly true in times of large transformational change such as the recent advances in artificial intelligence (AI) and GenAI. As companies accelerate their adoption of AI in multiple aspects of their business, they will also need to think through, manage, and reinforce new ethical challenges.

We may yet be far removed from a dystopian Skynet and Terminator-like threat of machines gone rogue. Still, understanding the implications of “human + AI” driven behavior and decision-making requires more upfront and explicit uncovering of hidden biases and other ethical considerations.

A Top Leadership Priority

More than half of respondents to a 2022 survey conducted by BCG for its 2022 Global ESG, Compliance, and Risk Report ranked business ethics among the top five topics most relevant to their compliance organizations. This was true regardless of industry or region.

However, even though companies across the spectrum agree that ethical behavior is a key priority, high-level principles are often abstract and can be difficult to operationalize, particularly in gray areas.

For example, the Sarbanes-Oxley Act, passed after the Enron scandal, requires public companies to disclose whether they have codes of ethics and whether any members of the senior management team are exempt. The Securities and Exchange Commission has gone further, requiring that listed companies have codes of ethics that apply to all employees, including senior management and directors.

And yet, as important as it is, a code of ethical conduct is far from sufficient to guarantee ethical behavior. While it serves as the guiding framework to outline principles and values, its effectiveness relies on a commitment (and sustained effort!) to implement and enforce it. Without effective processes, defined roles, a strong culture, etc. that permeates throughout an organization, a code of ethics can become a standalone document, disconnected from day-to-day operations. Ethical behavior must be ingrained in every aspect of an organization, reinforced by leadership and promoted at all levels. It requires ongoing training, communication, and a proactive approach to identify and address potential ethical dilemmas.

Ethical culture–and the mechanisms that support it–influences how employees interact, how decisions are made, and how a company engages with stakeholders. It goes beyond compliance and risk management to guide behavior in the face of unforeseen challenges—and when no one is watching.

Navigating Complex Terrain

It is often easier to define ethical behavior by its lapses, those high-profile public failures that often stem from corporate leaders intentionally skirting the rules or breaking the law. The spectacular collapse of Enron in 2001 is synonymous with ethical breakdown, as corporate leaders’ efforts to hide financial losses using off-the-books accounting schemes led to corporate bankruptcy and conspiracy and fraud convictions for those in charge.

While it may seem obvious that it was unethical for Enron’s leaders to lie to federal regulators 20 years ago, what about today, when accounting results may be inaccurate because an accounting algorithm uses an incorrect formula? Are corporate leaders in those instances also engaged in unethical behavior, enabling it, or–just as concerningly–completely unaware of the risks?

These are tough questions, and the discussion is both important and ongoing. Yet, in addition to new ethical challenges posed by technological advances, companies today also face more public scrutiny than in decades past. In an era of global hyper-transparency, instant judgments, and (public) whistleblowing, the public, including activist workers, is increasingly–and quickly–holding companies accountable for perceived ethical failings. An organization that has made clear its commitment to embracing a culture of ethical behavior and decision-making and put in place the necessary processes, tools, and culture to support such a culture will ultimately be better positioned to defend and weather criticism.

We want to highlight two specific and concrete ways that encouraging ethical behavior in the workplace can be enhanced in the context of AI: detecting bias to ensure fairness and ensuring corporate culture promotes transparency and learning across silos and levels, enabling employees to incorporate responsible AI in their day-to-day work.

In the realm of AI, detecting bias is a paramount task that is both technical and ethical. AI systems must be rigorously examined for biases that may be present, whether in underlying data, algorithms, or decision-making processes. Healthy discussion and debate within organizations are also essential in this regard, as they facilitate the identification and mitigation of biases. Transparency is the cornerstone of building trust into design, and is achieved by acknowledging and addressing biases, sharing insights into decision-making, and disclosing data sources. Learning from these discussions and experiences should be an ongoing process for organizations, leading to continued improvement. Achieving these learnings is impossible without weaving ethics into the overall organizational culture.

Identifying Hidden Biases

Everyone has cognitive biases or blind spots, and it is difficult for people to see their own potential for ethical failures clearly. Helping employees learn how to identify their biases and know when to ask for help is crucial.

Yet, in the AI space, hidden biases can emerge in lines of code that may have unintended and unwanted consequences. Historical data used to train algorithms is often rife with hidden biases, surfacing only when outcomes are closely tested. We’ve already seen this, for example, when a lender’s algorithm approves lower loan limits for women than for men or an algorithm for medical interventions leads to worse outcomes for Black patients.

How should workers who are responsible for coding these algorithms go about finalizing them? What tests should be run and when? How can bias in training data be uncovered? And what should a governance structure look like, and who should make the call on whether bias is sufficiently mitigated? Importantly, who is ultimately responsible for the consequences of that decision?

Potential solutions to address these questions are emerging: for example, model validation and bias testing are common practices that are already being applied to certain industries using traditional AI. Many organizations already rely on a system of checks and balances for high-impact decisions. Governance frameworks can be strengthened and include new experts and non-traditional voices.

Promoting Transparency, Healthy Discussion, and Learning

Ethical behavior starts at the top, and setting the right tone is critical. Senior management must walk the talk and lead by example. This means demonstrating commitment and support for a culture of ethics and communicating that across the organization.

In an environment where people feel comfortable discussing the issues related to ethics, bias, and AI broadly, they are empowered to speak up and ask questions without fear of retribution. Companies need to set clear expectations for ethical conduct and, crucially, make sure there are processes in place to support them. If people want to surface concerns about an ethical lapse–or even a perceived or suspected ethical lapse–they must know that there are procedures in place to allow that to be done safely and without fear of retribution. This means managers need to incorporate discussions about ethical considerations at every stage and communicate clear reporting paths for individuals to bring concerns to leadership. And then leaders must visibly react and learn from the ensuing discussions, making the processes and strategies better and rewarding those that identify potential issues and ways to improve. To mitigate risks and harms, companies should not operate in a vacuum. Instead, they should embed ethics into AI development, deployment, and monitoring–and engage in the ongoing interaction and collaboration required between technologists, businesses, policymakers, and society.

Risk and compliance professionals play a pivotal role in supporting and nurturing organizational culture. As AI’s influence grows, they must not remain on the sidelines but rather take a proactive and integrated approach to ensure ethical and responsible AI adoption throughout organizations. They must evolve by acquiring new skills, embedding themselves into businesses, and cultivating awareness throughout those organizations in order to keep pace with the speed of change brought on by AI integration. By providing guidance, fostering awareness, and advocating for ethical behavior, these professionals hold organizations accountable to the ethical values at the core.

For questions, concerns or more information about ACI Insights, please contact:
Chris Corbin
Associate Director of Marketing
American Conference Institute | The Canadian Institute | C5
E: [email protected]