AI, or models of computer machine-learning,
may piggy-back on ongoing risks to humanity, such as from nuclear bombs,
cyber-attacks, and biological weapons, including unleashing new pandemics, to render
any one of them more likely. That firms in the private sector develop advanced models
without a financial incentive to protect the public good exacerbates the risk.
To be sure, tech-firm executives were sounding the alarm most significantly in
2026. However, in reaching out to religious scholars and clergy for insight on
the ethics of AI in order to make the models ethical so their output would
be only ethical, certain executives incorrectly conceived of religion in terms
of ethics, and thus its criteria rather than those that are innate to religion.
This category mistake could result in advice that results in unethical computer
outputs. Anthropic is a case in point.
The full essay is at "Making AI Safe for Humanity."
