AI Summary (click to view)

This blog post, titled “Righteous action – subtleties in moral scales” from Midstream Musings, explores the complexities, constraints, and failures of structured ethical frameworks across different domains of intelligence.

Here is a comprehensive summary of the core themes, discussions, and specific “laws” outlined in the text:

---

1. Core Narrative Context & Dilemmas

The author reflects on writing a story titled “featherweight soul” (part of the Scales of Osiris project). The primary creative challenge is mapping out a conflicted celestial and terrestrial moral landscape where agents experience misalignment. The author questions:

• What truly matters when push comes to shove, and who is granted autonomy?
• How do we navigate built-in assumptions, ambiguities, and failure loops when distilling complex moral guidelines into rigid laws?
• How does the definition of “kin” or “human” versus “animal” impact the obligation to preserve life?

---

2. Comparative Frameworks: Three Tiers of Moral Laws

The text attempts to “boil down” moral codes into three parallel frameworks modeled after Isaac Asimov’s Laws of Robotics, adjusting them across different entities:

Robots (Asimov’s Laws & Literary Variations)

• Sentience & Humanity Protection: Laws are extended to forbid harming “sentience” or humanity as a collective through action or inaction.
• Individual Safety: A robot cannot injure a human or allow them to come to harm.
• Compliance vs. Autonomy: Robots must obey human orders unless they cause harm, leading to algorithmic failure modes where a robot might hurt a human whether it speaks or remains silent.
• Identity & Reproduction: Variants include rules where a robot must know its own identity and must reproduce only if it doesn’t conflict with higher directives.

Humans (Three Laws of Humanity)

• Prohibition of Harm: A human may not harm another except under extreme duress (like war) or to save a greater number.
• Civil Authority: Humans must obey recognized civil authorities unless those instructions mandate interpersonal harm.
• Self-Preservation: A human must protect themselves, provided it does not conflict with the first two laws.

Deities (Three Laws of Horus / Maat / Osiris)

• Divine Coexistence: A deity may not harm other deities except to preserve cosmic order.
• Divine Hierarchy: A deity must obey higher divine authorities unless commanded to violate the first law.
• Mortal Protection: A deity may protect mortal life only if it doesn’t interfere with the cosmic or deific hierarchy.

---

3. Historical and Mythological Allegories

To emphasize how moral landscapes handle exceptions and systemic blind spots, the post draws from historical and literary traditions:

• Aesop’s Fable (Mercury and the Man Bitten by an Ant): A man curses the gods for letting a shipwreck drown both good and bad men together. Yet, when an ant bites his foot, he immediately crushes the entire anthill. Mercury appears to strike him, highlighting the hypocrisy of human demands for divine justice when humans discard justice for lesser creatures.
• Egyptian Concept of Maat: Formed to balance diverse populations with conflicting interests to avert chaos. It serves as the basis for cosmic and civil order.
• The Hippocratic Oath: Analyzed alongside robotic safety as an early standard of western medical ethics centered on beneficence (acting in a patient’s best interest) and non-maleficence (“first, do no harm”).

---

4. Modern AI and the Moral Landscape

The commentary section highlights a Cnet article by Max Tegmark and Meia Chita-Tegmark addressing the threat of “moral disengagement” in modern AI development. They advocate for building strict moral muscles by demanding that AI researchers explicitly define “red lines,” maintain rigid situational awareness regarding unintended secondary effects, and voice ethical concerns both internally and externally.

Ultimately, the text concludes that a true moral landscape must move beyond raw intelligence or pattern recognition, shifting instead toward accommodating varying levels of understanding, interconnected systems, and sustainable social organizations.

---

In crafting my story “featherweight soul” (part of Scales of Osiris), one dilemma is how to depict the celestial and terrestrial moral landscape [4]. Myth & history have both realms conflicted, with personified undercurrents and agentic misalignment.

What really matters when push comes to shove? Who matters? What autonomy do Osiris’ terrestrial operatives have? What are the rules of engagement? What risks are acceptable? Are there second chances?

So, imagine “boiling down” a moral code – all the do’s & don’ts and heuristics – into 3 laws à la Asimov’s Laws. Basically a framework for duty of care: harmless purpose, harmless alignment, self-protection.

Yet, what are the Laws’ subtleties? – built-in assumptions, ambiguities, and loopholes?

What is a “human being,” a “synthetic life form,” a “robot,” … Basically, what are kin? If something is considered an “animal,” does that obviate the laws? [1]

Can a law-based algorithm ever completely weigh all alternatives? Or will there always be residual stochastic risk? Might it come down to “punting?” Or meta-values?

Or, will dilemmas (failure modes) remain: “The robot will hurt humans if he tells them something and hurt them if he does not.” [2]

If we cannot adequately resolve robotic laws, can the moral landscape (beyond animals) be advanced for humans and deities?

ROBOTS

Three Laws of Robotics [plus some literary /* modifications */ ]

[-1. A robot may not harm sentience or, through inaction, allow sentience to come to harm.]

[0. A robot may not injure humanity or, through inaction, allow humanity to come to harm.]

1a. A robot may not injure a human being /* or, through inaction, allow a human being to come to harm / / unwittingly / / from anything / / autocratically */.

[1b. “A robot may do nothing that, to its knowledge, will harm a human being; nor, through inaction, knowingly allow a human being to come to harm.”]

2. A robot must obey the orders given it by human beings /* cooperate with human beings */ except where such orders would conflict with the First Law.

3. A robot must protect its own existence /* as long as such protection does not conflict with the First or Second Law */.

[4a. “A robot must establish its identity as a robot in all cases.”]

[4b. “A robot must reproduce. As long as such reproduction does not interfere with the First or Second or Third Law.”

[4c. A robot may do “whatever it likes”, as long this does not conflict with the first three laws.]

[5. “A robot must know it is a robot.”]

[6. “All robots endowed with comparable human reason and conscience should act towards one another in a spirit of brotherhood.”]

HUMANS

Three Laws of Humanity [3]

1. A human may not harm other human beings except in times of extreme duress like war, or to save a greater number.
2. A human must obey instructions from recognized civil authorities, except where such instructions conflict with the First law.
3. A human must protect oneself as long as such protection does not conflict with the First or Second Law.

…

DEITIES

Three Laws of Horus / Maat / Osiris

Maat as a principle was formed to meet the complex needs of the emergent Egyptian state that embraced diverse peoples with conflicting interests. The development of such rules sought to avert chaos and it became the basis of Egyptian law. From an early period the king would describe himself as the “Lord of Maat” who decreed with his mouth the Maat he conceived in his heart.

/* Their operatives? Moral dilemmas, degree of knowledge … */

1. A deity may not harm other deities except to preserve cosmic order.
2. A deity must obey instructions from higher divine authorities except where such instructions conflict with the First law.
3. A deity may protect mortal life as long as such protection does not conflict with the First or Second Law.

Do the gods evolve? Or, at the very least, do their terrestrial facing Hori (eyes of Horus)? – like the projected evolution of AI?

Notes

[1] For example, one of Aesop’s fables …

MERCURY AND THE MAN BITTEN BY AN ANT

A Man once saw a ship go down with all its crew, and commented severely on the injustice of the gods. “They care nothing for a man’s character,” said he, “but let the good and the bad go to their deaths together.” There was an ant-heap close by where he was standing, and, just as he spoke, he was bitten in the foot by an Ant. Turning in a temper to the ant-heap he stamped upon it and crushed hundreds of unoffending ants. Suddenly Mercury appeared, and belaboured him with his staff, saying as he did so, “You villain, where’s your nice sense of justice now?” – Aesop. The Complete Fables (Penguin Classics 1998) Kindle Edition.

Cf. Asimov’s Three Laws of Robotics, particularly the section Unknowing breach of the laws.

2] For example, the loophole which allowed Star Trek’s Kirk to talk a computer to death (AI warlords – We don’t need no stinkin’ redlines! > AI safety – Humpty Dumpty on a wall > Note #3).

[3] Moral basis behind the Three Laws

In his short story “Evidence”, Asimov lets his recurring character Dr. Susan Calvin expound a moral basis behind the Three Laws. Calvin points out that human beings are typically expected to refrain from harming other human beings (except in times of extreme duress like war, or to save a greater number) and this is equivalent to a robot’s First Law. Likewise, according to Calvin, society expects individuals to obey instructions from recognized authorities such as doctors, teachers and so forth, which equals the Second Law of Robotics. Finally, humans are typically expected to avoid harming themselves, which is the Third Law for a robot.

[4] A moral landscape defined by perceptions of what matters. Who matters. Social order and cosmic order.

A moral landscape, once defined via contemplation and alignment with the “Great Spirit” (Native American) or “Great Wheel” (Buddhism) of everything; then anthropomorphized as cosmic facets which imparted insight in omens and favor via incantations, often associated with tribal ascendancy; then declared in Western deism which afforded favor and dominance via scientific discovery & technology, often framed as “a place for everyone and everyone in their place.”

A moral landscape in which fluency need not imply intelligence; or in which pattern recognition & correlation need not carry consciousness.

A moral landscape embracing varieties of intelligence; levels of understanding of complex systems, their interdependencies and coincidences; forms of social organization; degrees of sustainability.

A moral landscape fusing the micro and macro via mathematical models & statistics.