God From the Machine: AI and the Future of Humanity

Foreign Affairs  |  Sebastian Mallaby

Pope Leo XIV issued Magnifica Humanitas, the first papal encyclical addressing artificial intelligence, to defend human dignity and urge global regulation against technological risks. The document asserts that computational systems merely imitate human cognition without possessing genuine understanding, conscience, or moral experience. However, rapid advancements in large language models and reinforcement learning—demonstrated by neural networks parsing semantic concepts and autonomous systems learning through iteration—challenge traditional philosophical distinctions between organic consciousness and machine intelligence.

Beyond theological debates, the encyclical warns that unregulated artificial intelligence risks worsening economic inequality, destabilizing labor markets, degrading environmental resources, and lowering the threshold for automated warfare. While the Vatican highlights crucial ethical vulnerabilities, effective governance remains difficult due to the technology's sweeping, multifaceted impact across global society. Advanced artificial intelligence capabilities continue to accelerate, forcing policymakers, technologists, and religious authorities into unprecedented philosophical and regulatory confrontations over the future of human autonomy.

Comment

Integrating self-learning algorithms modelled on DeepMind’s Alpha-Zero into military command structures shifts tactical decision speed beyond human cognitive thresholds. Algorithmic autonomy compresses sensor-to-shooter timelines while obfuscating legal accountability when kinetic engagements execute without direct human intervention. This operational compression risks automated escalation cycles during high-intensity peer conflicts, as defensive systems react to machine-generated targets faster than command authorities can intervene.

Strategic Question for Discussion

If autonomous systems built on DeepMind's Alpha-Zero architecture are deployed to automate sensor-to-shooter targeting, which command-and-control mechanism best prevents unintended escalation during rapid multi-domain engagements?

Share your assessment in the comments below.

To read in details click here