As AI technologies advance rapidly, traditional frameworks for ensuring safety and ethics are proving insufficient. Reggie Townsend, head of AI ethics at SAS, argues for adopting a 'duty to care' approach that transforms governance into an ongoing commitment rather than a compliance exercise.
- Duty to care reframes AI governance as a continuous ethical practice.
- Traditional duty of care is compliance-based and backward-looking.
- Safety in AI involves enabling vulnerable self-expression, not just legal rules.
What happened
At the SAS Innovate on Tour event, Reggie Townsend introduced the concept of 'duty to care' as essential for safe and trustworthy AI. Unlike the conventional 'duty of care,' which focuses on legal liability and risk management, the duty to care is about embedding caring judgment directly into AI governance. Townsend created his role to lead AI ethics, governance, and social impact at SAS, engaging with regulators, customers, and the wider community to encourage a collective responsibility for AI development.
This shift emphasizes a governance model that integrates care as a foundational posture rather than a final box to tick. Townsend highlights how current governance models often feel like burdensome restrictions rather than tools for responsible innovation. Instead, he proposes care as the scalable element of human judgment and ethics, ensuring AI systems are built with empathy and purpose from the outset.
Why it matters
The context for Townsend’s argument is a world where change driven by AI outpaces institutional safeguards, fostering widespread uncertainty about safety and accountability. Traditional governance frameworks struggle to keep up with generative and agentic AI’s rapid deployment into society, frequently rewriting social norms only after issues emerge.
By advocating a duty to care, Townsend addresses the root of that instability: the tendency for regulation and ethics to be reactive and externally imposed. Embracing care turns safety into an active, shared responsibility across developers, industry players, regulators, and individuals. This approach recognizes that privacy and safety are about enabling individuals to control their exposure and expression in a complex digital age.
What to watch next
Industry and regulators around the globe will need to evaluate how to operationalize a duty to care within AI development and governance frameworks. This involves moving beyond legal compliance toward cultivating cultures of ongoing ethical reflection, care, and human-centered judgment. How companies like SAS influence broader regulation and industry norms will signal the direction of this evolution.
Additionally, responses to privacy and safety challenges from major tech firms will be a critical area to observe. Townsend points to Meta’s approach of pushing technological boundaries first and adjusting after public pushback as emblematic of the current reactive dynamics. The adoption of a duty to care ethos could drive a more proactive, balanced approach that aligns innovation with societal values from the start.