When it comes to aligning artificial intelligence with values, AI is the genie out of the bottle. For a refresher, the three wishes story is a cautionary tale. Ask for virtuous things and open a whole new amazing world. In the case of AI, wish for the wrong thing and it’s time to ask Alexa to page Sarah Connor (bonus points knowing the Terminator 2 reference without asking Claude).
Aligning AI use with values takes more than wishful thinking. It takes thinking through what is asked of AI and how it’s used. We asked members of the Fast Company Impact Council how they align AI with their values. Read what 13 of them shared on how to keep AI firmly in the good wishes camp.
1. KEEP VALUES IN CODE
We keep our values in code, not in prompts. The things we value most (what gets blocked and what requires human sign-off) run as a deterministic check that executes identically every time. A person approves anything irreversible. If your artificial intelligence alignment strategy is a well-worded instruction, you don’t have a strategy, you have a suggestion — Usman Khan, ADP
2. START WITH PURPOSE QUESTIONS
We start with purpose, not technology. Every artificial intelligence decision is measured against a simple question: Does it expand human potential, strengthen trust, and create healthier communities? We have governance, human oversight, and clear ethical guardrails, but our greatest safeguard is remembering that AI should augment and not replace human judgment, empathy, and accountability. Innovation only creates lasting value when it advances both performance and people. — Jean Accius, CHC: Creating Healthier Communities
3. ESTABLISH AN AI INNOVATION LAB TO TEST TOOLS
Our approach to artificial intelligence reflects our value of curiosity by encouraging exploration, learning, and innovation. We’ve established a data board to prioritize our efforts and an AI innovation lab where practitioners test emerging tools before they are introduced more broadly. Rather than waiting for perfection, we are adopting tools when they are “good enough” and making improvements along the way. We are also implementing policies and safeguards to protect both firm and client intellectual property. We may hit some bumps along the way, but we believe this approach is the best way to stay at the forefront of this rapidly changing technology. — Troy Thompson, SmithGroup
4. SHAPE THE PATH FORWARD
So much of artificial intelligence right now is operational enablement. It’s a tool and we use it. We are not the market force driving AI forward or slowing it down. That’s not our leverage. What we can do is help shape the path it takes and be deliberate about where we can most effectively do that for children. That means showing up in the rooms where AI is designed, shaping where it’s deployed, and developing use cases built with children in mind. — Michele Walsh, UNICEF USA
5. GIVE HUMANS THE FINAL WORD
We treat artificial intelligence as leverage, not a replacement for judgment. In advocacy, trust is the entire business. We use AI to research faster and draft smarter, but a human always owns the final word and the relationships behind it. We’re explicit with our team about where AI helps and where it can’t go, especially on anything touching a client’s reputation or sensitive policy. Values don’t live in the software. They live in the guardrails you build around it. — Howard Franklin, Ohio River South
6. UNDERSTAND HOW AI WORKS AND ITS LIMITS
I try to be intentional about when artificial intelligence is the right tool and when it isn’t. That starts with understanding how it works and where its limits are. The simplest test: Is the output the point, or is the work the point? Nobody sends someone else to the gym for them. The value is in doing the workout, not having it done. Some tasks are like that: The thinking is the point, and handing it to AI cheats you out of exactly what you needed. For everything else, the output is what matters, and AI is the obvious choice. The discipline is noticing when you’ve slipped into autopilot and being willing to zoom out and reset. — Blake Brannon, OneTrust
7. ASK: DOES IT BUILD CAPACITY OR REPLACE JUDGMENT?
We made a deliberate decision early on: Artificial intelligence enters our work through us, not around us. Every tool we adopt must answer one question: Does it build capacity or replace judgment? Conflating the two is where most organizations go wrong. We test AI against real work scenarios, not vendor demos, and document what works, what misleads, and where human judgment is irreplaceable. Values alignment isn’t a policy. It shows up in who’s in the room when the decisions get made. — Kellie Lauth, MindSpark
8. DON’T OUTSOURCE CORE COMPETENCIES
Our experienced professionals often use artificial intelligence as a tool to do more mundane things they would otherwise have done themselves, only more quickly. When we apply it to creative works like visual design and writing, it is for inspiration and not for generating finished products. We believe humans still have the upper hand in knowing what to design and build. That is a core competency of our team. We don’t want to outsource that to AI for the same reasons we don’t outsource those core competencies to other teams. — Brad Weber, InspiringApps
9. SHAPE POLICY UPFRONT, NOT AFTER
I make sure I have a seat at the table when decisions about artificial intelligence use are being made and not just reacting after they’ve already impacted our stakeholders. As communicators, we’re often the arbiters of truth for our organizations. So I push to help shape policy upfront instead of managing the fallout later. Being an active participant, not a bystander, is how you keep AI use aligned with your values. — Megan Noel, Covista
10. START WITH CUSTOMER NEEDS
My philosophy is that innovation must ultimately serve the customer. Our mission is to unlock financial progress for everyday people. That means we start with our members’ needs, not the technology. With artificial intelligence, it’s important to not get distracted by shiny objects or add technology for technology’s sake. We focus on applying it where it solves meaningful problems, improves the member experience, and helps us earn and strengthen trust. — Vineet Mehra, Chime
11. APPLY A PEOPLE-FIRST PACT
Our “people pact” is the test we apply to every artificial intelligence investment: Does it help people do their best work, live their best lives, and fulfill our purpose together? When AI creates real capacity for deeper thinking and better judgment, it passes. Wisdom and human judgment will never be automated. Companies that nurture both are the ones that outperform and outlast. — Jacqui Canney, ServiceNow
12. FOCUS ON COMMITMENTS TO CUSTOMERS
We view artificial intelligence as a tool to enhance human expertise—not replace it. We prioritize responsible, secure use of AI to improve efficiency, support better decision-making, and enable our teams to focus on higher-value work, while maintaining human oversight and protecting customer and company data. Any implementation must align with our commitment to safety, security, operational excellence, and delivering value to our customers, partners, and stakeholders. — David Klanecky, Cirba Solutions
13. ASK: DOES IT MAKE LIFE EASIER OR JUST SERVE THROUGHPUT?
I hold AI use against one primary test. Does it make people’s lives easier, or does it just serve throughput? If it doesn’t pass the test, it shouldn’t scale. — Todd James, Aurora Insights LLC