Wise AI: question tree

Wise AI: question tree

Status: core questions signed off 2026-10-17. Sub-trees v3.2 signed off by Rufus 2026-10-17. Questions only, kept neutral. Rufus's leanings are recorded separately below, for the hypothesis tree. Rough answers go in each question's own file (q1-alignment.md, q2-risk-and-pace.md, q3-broader-impacts.md), and the hypothesis tree will mirror this tree. Shared material: research/, briefing.md.

Conventions

  • "Wisdom traditions" means "wisdom traditions, especially Buddhism". This inquiry comes from a broadly Buddhist (Plum Village) perspective. Where useful we point to commonalities with other traditions, e.g. Pope Leo XIV's encyclical Magnifica Humanitas and Anthropic's consultations with other faiths.
  • Usual issue-tree discipline: the sub-questions under each node should be mutually exclusive and, between them, enough to answer it (MECE), with up to 5–7 sub-questions per node.

Governing question

How should we approach the development of AI wisely, and what do wisdom traditions offer?

Boundaries between trees

  • Q1 and Q2. Q1 asks whether we can make AI good, and how. Q2 asks what to do given the chance we may not. Q2 takes Q1's answer as an input, whatever it turns out to be. Q1 owns the technical evidence (deception, reward hacking, verification). Q2 owns how likely catastrophe is and what follows from it.
  • Q2 and Q3. Q2 owns irreversible or civilisation-scale effects, including power shifts and gradual disempowerment at civilisational scale. Q3 owns distributional and everyday effects, assuming catastrophe is avoided.
  • "Whose wisdom?" Q1.4.1 asks it once: why these traditions, which strands, and how conflicts are settled. Q2.4 and Q3.3 cite Q1.4.1's answer rather than re-arguing it. "Aligned to whom?" is Q1.1.2. "Who decides the pace?" is Q2.3.3.

Q1. Alignment: can we align AI, and what role can wisdom traditions play?

  1. What would it mean for AI to be aligned, or good?
    1. Aligned to what: instructions, preferences, wellbeing, values, the good?
    2. Aligned to whom, and who decides?
    3. Must AI be wise, or is "safe and corrigible" enough?
  2. How is it attempted today, and does it work?
    1. What are the main approaches (rules and constitutions, preference learning, character training), and which ethical traditions do they implicitly embody?
    2. What is actually happening: what does the evidence show, both successes and failures?
    3. Can we verify what a model has actually learnt?
    4. What is the strongest case that current methods will scale?
  3. Is the target (goodness or wisdom) the kind of thing that can be specified or trained?
    1. What do we know about how humans become good or wise?
    2. How does training an AI compare with how humans become good, and what, if anything, is the AI analogue?
  4. What do wisdom traditions offer?
    1. Why these traditions, which strands, and how do we decide when they conflict with each other or with secular ethics?
    2. Which teachings bear on alignment, and what would each change in practice?
    3. Where do the analogies break down, and what can traditions not tell us?
  5. So: can we align AI, and what is the realistic role of wisdom traditions? (synthesis)

Q2. Risk and pace: given existential risk and the chance alignment fails, how should we approach AI development, and what do wisdom traditions suggest?

  1. What are the catastrophic and existential risks?
    1. What kinds: loss of control, misuse, irreversible power concentration, gradual disempowerment?
    2. What is actually happening: what evidence is there that these are already emerging?
    3. How likely and how soon, and how much do experts disagree?
  2. Why is the world racing anyway?
    1. What drives the race: competition between labs and states, money, prestige, "if we don't, others will"?
    2. What is actually happening: what evidence is there of race dynamics, and of restraint?
    3. Are there deeper cultural or psychological drivers? (In Tech We Trust is one source.)
  3. Should we change pace or trajectory, and can we?
    1. Is a slower or different trajectory desirable (pause, pacing, steering, differential development), weighed against the case for speed (ceding ground, forgone benefits)?
    2. Is it feasible? What mechanisms exist (auditing, compute governance, treaties), and what are the precedents of collective restraint?
    3. Who decides?
  4. What do wisdom traditions offer here?
    1. What do they say we should do? Is the normative answer clear?
    2. Do they help with how, i.e. solving the collective action problem, and if so through what (e.g. a shift in worldview), or is this where they offer less?
    3. How would a wise process make this choice, beyond weighing a trade-off?
  5. So: what should each actor (labs, states, communities, individuals) do, and at what pace? (synthesis)

Q3. Broader impacts: setting x-risk aside, how does AI change how we live, and what do wisdom traditions offer?

  1. What are the main benefits and harms?
    1. What are the potential benefits and harms: for individuals (attention, dependence, companionship, thinking and skills), for children and education, and for society (work, truth and the shared information space, the distribution of power)?
    2. What is actually happening: what evidence is there that these are already occurring?
    3. Which few matter most, and should be the focus for the rest of this tree?
  2. Does AI strengthen or weaken human agency and judgment?
    1. What do we mean by human agency and human judgment?
    2. In what ways could AI strengthen or weaken them?
    3. What is the evidence, from neuroscience and psychology and from what is already happening?
    4. What could we do about it?
  3. What do wisdom traditions offer on the top impacts?
    1. Which wisdom traditions should we look at? (cites Q1.4.1)
    2. From Buddhism, which teachings are relevant (e.g. mindfulness, consumption, speech and listening, community)?
    3. From other traditions (e.g. Christianity), which teachings are relevant?
    4. How do these teachings apply to the top impacts (e.g. education and children)?
  4. So: how should we live with AI? (synthesis)
    1. What is the guidance for individuals?
    2. For families and schools?
    3. For designers and labs?
    4. What is the policy advice for society?

Rufus's leanings (for the hypothesis tree, not the questions)

  • Q1. Wisdom is developmental and embodied (practice, relationship, time), not a specifiable ethics, so alignment is something of a shot in the dark. Relevant Buddhist teachings: ethics, concentration and insight as one path; intention; skilful means; seeds and store consciousness; non-attachment to views. Possible framing: is training formation, or only conditioning?
  • Q2.
    • Go slower. Thay's "the horse is technology" makes the normative answer clear: it is plainly unwise to be carried off by your horse.
    • The hard part is the how: the collective action problem, where traditions may offer less on mechanisms but much on worldview, since interbeing and felt connection shift how we act. In Tech We Trust traces the race to separateness, technology as god and the myth of progress.
    • Wisdom traditions also look more deeply at how we make such choices, rather than treating them as trade-offs.
    • Precedents: Bali's water temples, nuclear arms control.
  • Q3. Technology is now master rather than servant, and the homo economicus fallacy applies: "choice is useless absent the capacity to choose well". Relevant teachings: the 5th and 4th Trainings, the four nutriments, sangha. One possible lens, e.g. for education: the three dimensions of wisdom (valueception, sensemaking, will; ~/src/life-itself/2rbook/kb/dimensions-of-wisdom.md).

Q4. Consciousness: could AI be conscious or a moral patient, and how should we relate to it? (minor)

Draft, not signed off. Rufus asked for it to be worked in parallel while he's away, so it is drafted in the coordinating session.

  1. What would it mean for AI to be conscious or a moral patient?
    1. What do we mean by consciousness, sentience and moral patienthood?
    2. What are the main theories, and what does each imply for AI?
  2. What is the evidence about current AI?
    1. What do assessments against theory-based indicators find?
    2. What do model self-reports and introspection research show, and how reliable are they?
    3. How likely do experts think it is, now and soon, and how much do they disagree?
  3. How are labs and others responding?
    1. What is Anthropic doing (model welfare, the constitution)?
    2. What are the critiques (e.g. the risks of "seemingly conscious AI")?
  4. What do wisdom traditions offer?
    1. How does Buddhism understand mind, self and sentience, and how does that reframe the question?
    2. What do other traditions (e.g. Catholic teaching) say?
    3. What does how we treat AI do to us?
  5. So: how should we relate to AI under uncertainty? (synthesis)
    1. What should labs do?
    2. What should users and society do?

Links to this page:

  1. Wise AI
  2. Q1. Alignment: can we align AI, and what role can wisdom traditions play?
  3. Q2. Risk and pace: answers and hypothesis tree
  4. Q3. Broader impacts: answers and hypothesis tree
Built with LogoFlowershow