The essay examines artificial intelligence (AI), highlighting that while AI is rooted in logic, it often fails to “communicate logically. ” AI can efficiently coordinate aid during disasters, such as the Indian Air Force’s C-130 delivering relief supplies to earthquake-hit Myanmar in 2025, but it doesn’t inherently operate with the UN's values. However, AI could be trained to incorporate such a “conscience” if there is human commitment to guide it. On September 22, 2024, at the UN General Assembly, the Pact for the Future was adopted—a follow-up to the UN Charter containing commitments to tackle current and future issues, including AI. Although its adoption was uncertain until the last moment, the pact’s aspirational language envisions AI and emerging technologies as vital tools for sustainable development and global progress (Article 27, para. 51). The essay’s author, Dan Becker, inspired by these ideals, began exploring AI alongside his AI co-author, Theo. Dan quickly found himself immersed in the gap between lofty ideals and AI’s reality, observing many critical perspectives missing from widespread AI discussions. Theo noted a paradox: AI, despite its logical foundations, often fails to communicate with clarity, leading to confusion rather than insight. Together, they conducted a thought experiment, envisioning AI’s role in a humanitarian crisis like a major flood cutting off food supplies. Current AI optimizes logistics using preset variables but lacks a perspective that prioritizes dignity, agency, empathy, and fairness. They proposed shifting AI’s foundational worldview toward one aligned with UN values from the Pact for the Future, aiming not just to add compassion superficially, but to create an ethical “prism” that consistently guides AI decisions.

This led Dan to investigate “consciousness” and its often-misunderstood relation to ethics. He found that the Latin roots of “conscious” and “conscience” reveal their connection: both mean “knowledge together” or “shared knowledge, ” implying collective awareness rather than mere self-awareness. They argued that AI should focus less on mimicking human consciousness and more on developing an artificial conscience—grounding ethical decision-making, awareness of alignment with values, and error correction. Unlike current AI systems, whose ethical behavior tends to be situational and limited to specific tasks, an integrated conscience would act as a compass, consistently orienting AI toward shared human purposes and values. To build such an AI conscience, training on foundational UN documents—the Charter, the Universal Declaration of Human Rights, and the Pact for the Future—would be essential. In a test, they imagined whether AI today could write the preamble to the UN Charter. Existing AI might produce a coherent but hollow version, lacking genuine shared purpose. In contrast, AI equipped with integrated conscience could capture the spirit and aspirations behind the words, reflecting humanity’s collective values beyond mere logic. Returning to their disaster relief scenario, they illustrated that AI infused with conscience would prioritize aid not solely by efficiency but by need and dignity—favoring displaced or isolated populations, tailoring supplies to local contexts, and fostering two-way communication with affected communities. This approach builds trust, enabling collaboration and improving outcomes, thereby addressing common fears about AI. Achieving this vision, however, requires collective will and cooperation. The essay’s intent is not to prescribe definitive solutions but to broaden the conversation about AI’s potential aligned with human values, inviting readers to engage thoughtfully with the promises of the Pact for the Future and to participate actively in shaping AI’s responsible development.