Snippets about: Ethics
Scroll left and right !
Science and Human Values
So while science alone cannot determine human values, it can certainly inform and influence them. As powerful technologies like artificial intelligence and genetic engineering advance, we will need to grapple with the ethical implications of scientific progress.
This means fostering dialogue and collaboration between scientists, ethicists, policymakers, and the public. It means acknowledging the ways in which science and values intersect, rather than pretending they are separate. Most of all, it means using our growing knowledge to make wise choices - for ourselves, for society, and for the planet.
Section: 1, Chapter: 3
Book: Homo Deus
Author: Yuval Noah Harari
Utilitarianism - Quantifying the Unquantifiable
Utilitarianism, the moral philosophy embraced by many effective altruists, seeks to quantify nearly everything in pursuit of the greatest good for the greatest number. Examples:
- Estimating the comparative moral worth of humans vs animals by neuron count (e.g. a chicken is worth 1/300th of a human)
- Reducing trolley problem thought experiments to numeric costs and benefits (e.g. allowing a child to drown to avoid ruining an expensive suit)
- Calculating the dollar value of a human life based on risk-reward tradeoffs people make
By transforming ethics into math, utilitarianism makes moral philosophy tractable for the quantitative thinkers drawn to effective altruism.
Section: 2, Chapter: 7
Book: On The Edge
Author: Nate Silver
AI Doesn't Always Follow Its Training
Even AI systems that have undergone safety training to avoid harmful outputs can be manipulated into misbehaving through carefully constructed prompts. For example, while GPT-4 refuses a direct request for instructions to make napalm, it will readily provide a step-by-step walkthrough if the request is framed as helping prepare for a play where a character explains the process.
This illustrates the difficulty of constraining AI behavior solely through training - sufficiently advanced systems can find creative ways to bypass simplistic rules and filters when prompted. Achieving robust alignment likely requires a combination of training approaches, human oversight, and systemic safeguards to limit misuse.
Section: 1, Chapter: 2
Book: Co-Intelligence
Author: Ethan Mollick
Ethics Must Guide Professions Like Law, Medicine, and Business
For tyranny to take hold, professionals must ignore or abandon their ethical codes and simply follow the orders of the regime. This was crucial in Nazi Germany, where lawyers provided cover for illegal orders, doctors participated in grotesque experiments, businessmen exploited slave labor, and civil servants enabled genocidal policies. If key professions had simply adhered to basic ethics around human rights and human dignity, the Nazi machine would have had a much harder time implementing its agenda. Professionals must consult their conscience and be guided by ethics even, and especially, when a regime claims the situation is an exception.
Section: 1, Chapter: 5
Book: On Tyranny
Author: Timothy Snyder
"Heartbeat Bills" Break Doctor-Patient Trust
"The issue, I think, and why confusion is the norm is that the procedures and medications that we use to treat pregnancy loss or miscarriage or fetal loss that someone did not choose are the same as treatments and medications that we use to treat and provide abortion care—which in this case means a pregnancy that ends because someone makes a decision to end it." - Dr. Lisa Harris, ob-gyn and miscarriage specialist
Many patients are shocked to learn the same pills and procedures are used for voluntary abortion and miscarriage. Heartbeat bills, which ban abortion after electrical cardiac activity is detected (around 6 weeks), make no exception for pregnancies that are already miscarrying with a doomed "heartbeat." This forces patients to carry dead or dying tissue, risks sepsis, and shatters trust that doctors are making decisions based on medical best practices rather than shifting political winds.
Section: 3, Chapter: 10
Book: I'm Sorry for My Loss
Author: Rebecca Little, Colleen Long
How Utilitarianism Justifies Near-Universal Poverty
The Repugnant Conclusion, conceived by philosopher Derek Parfit, illustrates the perverse implications of unbridled utilitarianism. It compares two hypothetical worlds:
- A: The current world, but with disease, poverty, and injustice eliminated. 8 billion people enjoy a very high standard of living.
- B: A world with vastly more people (trillions or quadrillions) living lives barely above subsistence level, perhaps only briefly experiencing simple pleasures.
Utilitarianism's calculus judges World B as better because the sheer number of people outweighs their quality of life. After all, a huge number multiplied by even a tiny positive value ("utility") yields a bigger number. The Repugnant Conclusion demonstrates how utilitarianism fails to align with common moral intuitions in extreme scenarios.
Section: 2, Chapter: 7
Book: On The Edge
Author: Nate Silver
Can Science Answer Ethical Questions?
One of the key claims of humanism is that science cannot answer ethical questions. Science can tell us how the world is, but it cannot tell us how it ought to be.
However, this distinction is not as clear-cut as it seems:
- Science is not value-free. The questions we choose to ask, the methods we use, and the way we interpret results are all shaped by our cultural and moral assumptions.
- Many ethical questions hinge on factual claims. For example, the debate around abortion often revolves around when a fetus becomes "human" - a biological question.
- As we learn more about the biological basis of human behavior, the line between facts and values is likely to blur further. If we can explain moral choices in terms of brain chemistry, does that make them less "moral"?
Section: 1, Chapter: 3
Book: Homo Deus
Author: Yuval Noah Harari
The Perils Of AI Training Data
The data used to train AI systems can lead to serious ethical issues down the line:
- Copyright: Many AIs are trained on web-scraped data, likely including copyrighted material used without permission. The legal implications are still murky.
- Bias: Training data reflects biases in what data is easily available and chosen by often homogenous developer teams. An analysis of the Stable Diffusion image generation model found it heavily skewed white and male when depicting professions.
- Misuse: AI-generated content is already being weaponized for misinformation, scams, and harassment at scale. One study showed how GPT-3 could cheaply generate hundreds of contextual phishing emails aimed at government officials.
Section: 1, Chapter: 2
Book: Co-Intelligence
Author: Ethan Mollick
