Big AI labs are hiring philosophers
Big AI labs are increasingly hiring academic philosophers to help shape model “constitutions,” align systems with ethical frameworks like deontology and consequentialism, and advise on questions of consciousness and societal impact. Commenters are divided over whether this reflects genuine concern for AI ethics or is mostly PR theater, noting that philosophers may end up legitimizing pre‑chosen business goals or performing low‑status work. Others point out that philosophy’s tools for clarifying concepts, reasoning about values, and questioning assumptions are well suited to emerging AI risks, even if such roles remain rare compared with traditional engineering positions.
Hiring philosophers for AI labs
- Several commenters with philosophy backgrounds ask how to pivot into these roles.
- Others respond that a minor is insufficient; roles likely require being well-published and highly cited.
- Competition is seen as intense: vague roles, very high compensation, and many similarly qualified candidates.
Skepticism about scale and motives
- Some doubt the article’s claim that labs are hiring “many” philosophers, suspecting only a handful exist alongside hundreds of engineers.
- There’s concern philosophers may be hired mainly for PR or to affirm pre-decided positions rather than critically challenge them.
- Analogies are made to nutritionists hired after a fast-food chain is built, or psychologists hired to make social media more addictive.
Philosophy’s role in AI behavior and ethics
- Discussion highlights two main ethical frameworks: deontology (rules, duties, constraints) and consequentialism (cost–benefit, utilitarian reasoning).
- Different labs are perceived as leaning toward one or the other in their “constitutions” and safety goals.
- Some see philosophers as helpful for value-specification, model training constraints, and questions about potential AI sentience and moral status.
- Others argue sociologists or mathematicians might be more appropriate for societal or technical issues.
LLMs, context, and “philosophical” prompting
- Multiple comments note LLMs perform better when given rich context: the problem, intent, and reasoning behind a feature, not just imperative instructions.
- Debate over whether this is truly “philosophy” or simply better problem framing; consensus leans toward it being context, though some see overlap with practical/philosophy-of-action style thinking.
- Parallels are drawn to human dev work and the “XY problem” of specifying solutions instead of underlying problems.
Consciousness, political theory, and alignment
- Thread explores thought experiments: trolley problems, graded consciousness from animals to AI, and whether language-only systems fit that spectrum.
- There is disagreement about how future superintelligent AIs might evaluate human political norms (e.g., secession, self-determination).
- Some point out that AI companies have incentives to insist models are not conscious, to avoid ethical constraints on usage.
Academia, careers, and trust
- Claims of a “haemorrhaging” of philosophers from academia are met with skepticism; philosophy posts remain scarce and competitive.
- Some academics are seen as well-suited to business-facing roles, others only to traditional scholarship.
- Broader worry emerges about a “negative-trust future” where PR, AI-generated content, and corporate narratives increasingly shape what the public believes about AI.