The Ethics of Abstention: Is It Unethical Not to Use Synthetic Intelligence?

We speak incessantly of “ethical AI,” yet rarely of the ethics of abstaining from AI. In an era where social systems grow ever more complex—interdependent supply chains, planetary health, cyber-security, energy transitions—one must ask: is it ethical to deny ourselves the disciplined use of a tool expressly designed to reason over complexity? Or, put differently: are we making our institutions AI-ready, or postponing adoption for fear of misuse—thereby accepting the very harms we claim to avoid?

 

 

A prudent society does not worship technology; it governs it. The global governance architecture has matured quickly. The EU AI Act—now in force with phased obligations for banned uses, general-purpose AI transparency, and high-risk systems—codifies a risk-based approach. This is not techno-optimism; it is accountable adoption with binding dates and duties.

 

Complementing regulation, the NIST AI Risk Management Framework and its 2024 Generative AI Profile offer operational scaffolding: governance, testing, incident disclosure, content provenance—concrete actions that organizations can implement today. Ethical readiness is no longer a slogan; it is a checklist.

 

At the multilateral level, the UN General Assembly’s 2024 resolution urges states to bridge digital divides and steer AI toward the Sustainable Development Goals, while warning against uses incompatible with human rights. The 2024 Seoul AI Safety Summit further pressed for interoperability among governance regimes, and cooperation among emerging “AI safety institutes.” The ethical direction of travel is clear: use AI, but use it well.

 

 

Human decision-making confronts bounded rationality—our cognitive limits when facing complexity. This is not an indictment of human dignity; it is a sober observation stretching from Herbert Simon’s classic work on complex systems to Rittel and Webber’s “wicked problems” in public policy. Wicked problems lack definitive formulations and elude “once-and-for-all” solutions; they require iterative sense-making over sprawling data. Refusing an instrument that enlarges our analytic horizon may amount to culpable negligence when stakes are societal.

 

Governments are already experimenting accordingly. OECD, World Bank and others document how AI can augment policy design, service delivery and crisis response—provided guardrails are in place. The question, then, is not whether AI belongs in the public square; it is whether our institutions can integrate it responsibly and equitably.

 

 

The 2025 Stanford AI Index reports rapid improvements in technical performance, broad enterprise adoption, and steep drops in inference cost—lowering barriers to responsible deployment across sectors. At the same time, governments accelerated policy activity and safety benchmarks—an institutional countermove that tempers capability with oversight. This dual motion—capability plus governance—is precisely what ethics demands.

 

But a nuanced point is essential: meta-analyses show that human-AI teams do not automatically outperform the best humans or the best AI in decision tasks; hybrids often shine more in creative work. The ethical implication is not to abstain, but to design collaborations deliberately: specify roles, calibrate trust, train users, and measure outcomes. Ethics here is engineering discipline, not myth.

 

 

Ethical use also requires honesty about externalities. AI’s compute intensity raises material concerns for energy and water use. Policymakers and developers must track the “AI footprint,” invest in efficiency, and align deployment with sustainable grids. Refusing to measure costs is as unethical as refusing to reap benefits.

 

Likewise, global equity matters. The UN resolution and Seoul Declaration emphasize inclusion, capacity-building, and cooperation. If advanced AI remains a privilege of a few, we reproduce digital dependency and fragility. Ethical AI is therefore a development agenda as much as a safety agenda.

 

Finally, human dignity must remain the axis. The Holy See’s recent guidelines on AI insist that technology complements, not replaces, human judgment—warning against dehumanization, environmental harm, and the erosion of responsibility. Abstention is not the path to dignity; stewardship is.

 

 

“Artificial” often suggests imitation; “synthetic intelligence” better captures systems that synthesize across modalities to generate genuinely novel outputs. The term has a lineage in philosophy of AI and contemporary policy writing, and clarifies our intent: to use machines not as simulacra of minds, but as constructive partners in scientific, medical, and civic reasoning. Naming matters because it shapes governance goals.

 

 

What, concretely, does “ethical to use” look like?

1.Adopt risk-based governance (EU/NIST). Map use-cases, test pre-deployment, document data and model lineage, and disclose incidents.

2.Design human-AI workflows for the task. Where accuracy trumps creativity, privilege decision-assistance with clear escalation; where ideation matters, leverage generative breadth. Measure outcomes, not intentions.

3.Invest in equity and resilience. Fund capacity in undeserved regions, ensure multilingual access, and couple deployments with compute-efficiency and clean-energy strategies.

4.Anchor dignity. Preserve human oversight, rights, and accountability; avoid applications incongruent with international humanitarian and human-rights law.

 

 

In our own work with the IASC Longevity by Design initiative and the forthcoming interventions connected to the Ukrainian Reconstruction Summit, we are exploring AI-assisted protocols that integrate neuroscience, regenerative medicine, and longitudinal monitoring to support post-traumatic recovery—for victims and first responders alike. This is emblematic of the thesis above: the ethically necessary use of AI to address suffering at scale, governed by strict clinical, legal, and spiritual safeguards.

 

⸻

 

Conclusion

 

To refuse synthetic intelligence because it is powerful is to mistake power for vice. Power without stewardship corrupts; power with stewardship heals. Our fear should no longer be that AI will act without ethics, but that we will fail to build the ethical, institutional muscles to wield it for the common good. The right question is not “Should we use AI?” but “How dare we not—when human life, peace, and the planet’s future require every responsible instrument of reason we can bring to bear?”

 

⸻

 

Selected references

•EU AI Act timeline & obligations (European Parliament brief; European Commission).

•NIST AI Risk Management Framework – Generative AI Profile (2024).

•UN General Assembly Resolution on safe, secure, trustworthy AI (2024).

•Seoul AI Safety Summit, Seoul Declaration (May 21, 2024).

•Stanford AI Index 2025 – adoption, costs, policy activity.

•Nature Human Behavior (2024) – meta-analysis on human-AI teams.

•OECD.AI – AI Compute & Climate (energy/water footprint).

•AP News – Vatican guidelines emphasizing human primacy and responsibility.

•IASC blog – Regeneration for Humanity / Ukrainian Reconstruction Summit (post-trauma protocols).

•On “synthetic intelligence” as a term: historical and policy usage.