“We need to ensure AI is serving people, not the other way around.”
At an international gathering of AI safety institutes, Gina Raimondo said, “We need to ensure AI is serving people, not the other way around.” The sentence provides a simple test for a complicated technology. AI should be judged by what it enables people and communities to do, not only by benchmark scores, investment, or the amount of work it can automate. A system serves people when it improves health, expands access to knowledge, removes dangerous or tedious burdens, strengthens creative and professional capacity, and operates under conditions that protect dignity, rights, and security.
The distinction matters in the workplace. Automation can help employees spend less time on paperwork and more time on skilled or relational tasks. It can also become a tool for surveillance, unrealistic quotas, deskilling, or abrupt displacement. Whether AI serves workers depends on who participates in adoption, how productivity gains are shared, and whether people receive training and real opportunities to move into new roles. Efficiency is not a complete measure of benefit. Quality, autonomy, safety, wages, and the resilience of communities belong in the calculation.
Human-centered development also changes how systems are designed. Developers begin with a real need, include domain experts and affected users, test accessibility and subgroup performance, and create clear routes for feedback. They limit data collection to what is necessary and keep people responsible for high-stakes outcomes. Government can support that approach through research, standards, procurement, education, and competition policy. Public institutions are not merely referees after harm occurs; they can shape markets toward applications that improve health, education, infrastructure, and scientific discovery.
Raimondo’s quote refuses technological fatalism. Society is not required to accept every possible use simply because it can be built, and it does not have to choose between innovation and human welfare. The direction of AI depends on incentives, rules, investments, and everyday design decisions. Asking whom a system serves reveals whether benefits are broadly shared or concentrated while costs are pushed onto others. The strongest future for AI is not one in which people reorganize their lives around machines. It is one in which powerful tools remain deliberately organized around human purposes. The service test is deliberately reusable. For any proposal, leaders can ask who gains time, health, knowledge, income, or agency—and who loses privacy, bargaining power, opportunity, or voice. Clear answers make it easier to improve a system, limit it, or decide that it should not be deployed.
A useful system should improve safety, knowledge, health, creativity, opportunity, or the quality of work in a measurable way.
Capability alone is not the outcome; the question is how people’s lives and choices change.
AI can remove routine burdens, but it can also intensify surveillance, pace, and job insecurity.
Worker participation, training, and fair sharing of productivity gains help distinguish augmentation from extraction.
Standards, procurement, research funding, education, and competition policy influence which applications become profitable and widespread.
Government can reward systems that solve public problems while limiting uses that undermine rights or security.
Explore more "Quotes of The Day"
Continue exploring how writers, thinkers, artists, and advocates have described the choices that shape an everyday life.