The rapid acceleration of generative and agentic artificial intelligence has introduced a profound sense of instability across the global political and social landscape. As technological capabilities outpace the institutions designed to manage them, a growing consensus suggests that existing systems may be insufficient to ensure public safety and ethical integrity. At the recent SAS Innovate on Tour event in Liverpool, Reggie Townsend, Vice President of Data Ethics at SAS, addressed these concerns by proposing a fundamental shift in how corporations approach the development and deployment of AI. Townsend’s framework moves beyond traditional compliance-based ethics toward a proactive "duty to care," arguing that governance must be woven into the fabric of innovation rather than treated as an external constraint.
The Linguistic Shift: From Duty of Care to Duty to Care
A central pillar of Townsend’s philosophy is the distinction between a "duty of care" and a "duty to care." In corporate law and fiduciary practice, a "duty of care" is a well-established principle requiring directors to act in the best interests of the firm and its shareholders. However, Townsend characterizes this as a backward-looking, externally imposed obligation—a checkbox exercise intended to avoid litigation or regulatory penalties. In this traditional model, AI ethics is often perceived as a "bolt-on" requirement or a "brake" on the speed of business.
Conversely, a "duty to care" represents a voluntary, forward-looking posture. Townsend argues that this shift in prepositions changes the nature of governance from a reactive gatekeeping function to a proactive exercise in judgment. "Governance to me should be an exercise in care," Townsend stated during the Liverpool summit. "It’s the way we scale our judgment, but a judgment, I think, should be rooted in care." By adopting this stance at the beginning of the development lifecycle, organizations can transform governance into a core component of the creative process, ensuring that ethical considerations are integrated into the Python code and flow diagrams of developers from day one.
Regulation as a Mechanism for Public Safety
The conversation surrounding AI regulation is frequently framed as a conflict between innovation and oversight. However, Townsend suggests that the public demand for regulation is essentially a demand for safety. He posits that while governments have a duty to protect citizens in exchange for tax revenue, private industry shares a reciprocal responsibility to provide products that do not jeopardize the well-being of their users.
This definition of safety extends beyond physical harm to encompass "digital safety"—the ability for individuals to express themselves and pursue opportunities without fear of exploitation or unauthorized surveillance. In this context, privacy is redefined not merely as the absence of data collection, but as a conversation about the range of exposure an individual chooses to accept. Townsend notes that the current "rhythm" of the industry involves companies pushing the boundaries of social norms until they meet resistance. He cited Meta’s experimentation with biometric recognition in smart glasses as an example of a firm testing the limits of privacy, only to retract features following public or journalistic pushback.
The Chronology of Ethical Integration at SAS
The establishment of the Data Ethics Practice at SAS was not the result of a top-down mandate but rather an internal initiative led by Townsend approximately four years ago. Recognizing that AI was advancing at a rate that could either expand opportunity or deepen historical harms, Townsend presented a vision to the SAS executive leadership team to formalize the company’s commitment to responsible innovation.
SAS Chief Technology Officer Bryan Harris noted that the proposal was accepted due to Townsend’s ability to articulate the intersection of technical necessity and social impact. Today, the practice reports directly to the CTO, a structural decision that grants the ethics team significant influence over product roadmaps and profit strategies. This reporting structure is rare in the technology sector, where ethics teams are frequently siloed within legal or human resources departments. Since its inception, the mandate has expanded from general data ethics to include AI governance, regulatory standards, risk intelligence, and accessibility innovation.
Moving from Human-in-the-Loop to Human-in-the-Lead
As AI agents increasingly interact with other automated systems, the risk of obscuring human agency grows. Townsend advocates for a transition from "human-in-the-loop" to "human-in-the-lead." This distinction emphasizes that while machines can compute and decide at rates far exceeding human capacity, the ultimate responsibility for the outcomes of those decisions remains with human actors.
To achieve this, SAS emphasizes the importance of AI literacy. Townsend argues that while not every citizen needs to become a data scientist, a baseline level of understanding is required to exercise proper judgment. He compares the introduction of AI to the introduction of calculators in schools: calculators are introduced only after a child has mastered basic arithmetic. Similarly, society must decide when and how AI should be permitted to interact with vulnerable populations, such as children, based on developmental readiness rather than technological availability.
Supporting Data and the Global Regulatory Landscape
The push for a "duty to care" arrives amidst a surge in global AI legislation. The European Union’s AI Act, which entered into force in 2024, represents the world’s first comprehensive horizontal regulation for AI, categorizing systems by risk level and imposing strict transparency requirements on "high-risk" applications. In the United States, the White House Executive Order on Safe, Secure, and Trustworthy AI (2023) and the NIST AI Risk Management Framework provide guidelines for federal agencies and private firms to mitigate algorithmic bias and security vulnerabilities.
Market data suggests that ethical considerations are becoming a competitive necessity. According to the 2024 Edelman Trust Barometer, public trust in AI has declined as the technology becomes more pervasive, with 35% of respondents expressing concern about the rapid pace of change. Furthermore, a study by Gartner predicts that by 2026, organizations that prioritize AI transparency and ethics will see a 25% improvement in user adoption and brand loyalty compared to those that do not. Townsend’s focus on "irresistible responsibility" aligns with these trends, suggesting that ethical integrity is not just a moral imperative but a prerequisite for market viability.
The Liverpool Football Club Parallel: Systemic Reform
During the Liverpool event, Townsend’s philosophy was contextualized through the local history of the city, specifically the legacy of Liverpool Football Club. The club’s history provides a poignant example of how a "duty to care" can lead to systemic reform. The 1989 Hillsborough disaster, which resulted in the deaths of 97 supporters due to overcrowding and police failure, was initially blamed on the fans themselves. It took nearly three decades of advocacy by the families of the victims to overturn that narrative and establish the truth of systemic failure.
The response to the disaster was the Taylor Report, which mandated the conversion of major English football grounds to all-seater stadiums. This was a fundamental change to the system itself, ensuring that the conditions for such a tragedy could never recur. Townsend draws a parallel here to the tech industry: rather than punishing individuals for making poor choices within broken systems, the goal should be to fix the system so that the responsible path is the easiest one to follow.
The memorial at Anfield, featuring an eternal flame and a statue of Bob Paisley carrying an injured player, serves as a reminder that a culture of care is built on acknowledging failures and prioritizing the well-being of the collective. For SAS and Townsend, this serves as a blueprint for AI governance. By casting responsibility in "bronze"—making it a permanent and visible part of the corporate identity—firms can build the trust necessary to sustain innovation.
Broader Implications and the Future of AI Governance
The implications of Townsend’s "duty to care" extend beyond the immediate concerns of data privacy and algorithmic bias. As AI systems become more autonomous, the boundary between personal, industrial, and governmental governance will continue to blur. Townsend argues that rules will inevitably run out precisely when they are needed most, making individual and corporate judgment the final line of defense.
"There are no absolute rights and wrongs," Townsend noted, using the example of a driver running a red light while rushing to the hospital. "Those are matters of judgment." In the realm of AI, this means that developers and executives must be empowered to deviate from standard protocols when the safety or dignity of a human being is at stake.
The success of the SAS model suggests that a profit-driven enterprise can successfully integrate deep ethical considerations without sacrificing its competitive edge. By treating ethics as a "value-add" rather than a "cost-center," SAS aims to demonstrate that a "duty to care" is the most sustainable path forward in an increasingly automated world. As other tech giants grapple with regulatory scrutiny and public skepticism, the transition from compliance-based "duty of care" to a proactive "duty to care" may become the defining characteristic of the next era of artificial intelligence.
