Good people refuse to do bad things
Engineers and tech workers are debating whether it’s ethical to stay in high-paying roles at companies building powerful AI systems they fear could be harmful, or even existentially dangerous. Some argue that “good people refuse to do bad things” and should quit rather than rationalize their participation, while others stress moral gray areas, structural pressures, and obligations to family that make such choices far from simple. The exchange broadens into questions about AI hype versus real risk, the limits of personal responsibility under capitalism, and whether meaningful change is more likely from insiders or from those who walk away.
Ethical responsibility vs. complicity
- Several recount refusing unethical requests (e.g., falsifying financial reports), stressing that you should be prepared to actually resign, not just threaten it.
- Others argue that many “bad” outcomes arise from ordinary people following orders or routines (“banality of evil”), not cartoon villains.
- A recurring theme: doing “small” compromises repeatedly erodes your capacity to care, spreading to teams, products, and users.
Moral complexity and “good people”
- Many think “good people refuse to do bad things” is too binary; real life is shades of gray and people often misjudge what’s good/bad.
- Some defend moral relativism (different cultures, contexts), while others push back with examples like slavery or child labor to show not all “grays” are equal.
- One view: you may not define “good” cleanly, but you can often identify concrete harms (coerced risk, job loss without warning, irreversible damage).
Money, power, and structural constraints
- Strong disagreement over “you need financial security to do the right thing.”
- One side: family, kids, visas, and mortgages heavily constrain people’s choices.
- The other: well‑paid tech workers in particular often hide behind this as an excuse.
- Some emphasize structural issues (shareholder primacy, lack of safety nets), others insist individual choices are still morally significant.
Working on AI and other contentious tech
- Some argue that if lab leaders publicly talk about non‑trivial extinction risk, continuing to build such systems is unethical, no matter the salary.
- Others are deeply skeptical of AI doom narratives, seeing them as quasi‑religious, Rationalist/EA‑driven, or opportunistic marketing by companies and ex‑employees.
- Debate over AI’s real risks: extinction vs. mundane harms like climate impact, surveillance, job loss, cyberattacks, and disinformation.
- A contrasting view: AI is just another powerful tool; “Pandora’s box” is open, so the choice is which faction you empower, not whether development happens.
“No ethical consumption” and limits of purity
- Some note that if everyone stopped doing anything tied to harm, the global economy would collapse; all consumption under capitalism is tainted.
- Others respond that this doesn’t absolve people from avoiding the worst roles (e.g., surveillance capitalism, online gambling, defense, frontier AI) or from being honest that they’re staying mainly for money.