Changes

How the published reasoning has changed, release by release. Earlier wording is kept.

Detailed release history

Open a release to inspect individual edits.

Expand all six trees from David’s analysis
25 October 2026 · 2026-10-25.2

25Oct 2026

The goal becomes substantive, competing explanations remain open, and conditional interventions follow the diagnosis. A candidate conflict, prerequisites and conditional steps complete the six views. All reasoning remains proposed; no evidence, selected setting, owner, date, trial or observed outcome is invented.

Reworded
Replace the initial procedural goal with David’s proposed substantive goal.
Before~~People make informed, revisable choices about developing and using AI.~~
Current wordingPeople and institutions approach AI development and use wisely, with defensible reasons for the aims, risks and responsibilities they accept.
People and institutions approach AI wisely, with defensible aims, risks and responsibilities.
Reworded
Retain uncertainty as one competing hypothesis and record that it could also encourage restraint.
and 1 connection
Reworded
Replace the direct connection with four joint conditions: review points, usable evidence, authority and incentives, and feasible alternatives.
Before~~Explicit review points alone were connected to consequential revision.~~
Current wording“Decision makers use explicit review points to revisit AI choices against emerging evidence.” and “Usable evidence about the decision and its consequences is available.” and “Decision makers have authority and incentives to act on review findings.” and “Feasible alternatives to the current decision are available.” causes “AI choices can be revised when evidence changes their justification.”
and 1 connection
Added
Incorporate the supplied analysis across all six trees, preserving its attribution and unresolved reasoning.
AI’s intended ends are defensible and its pursuit of them is adequately supported.Development choices account for alignment uncertainty, catastrophe and alternatives.Adoption decisions account for agency, relationships and institutions.Treatment of AI responds responsibly to uncertain moral standing.Decision makers can restrain action before consequences become irreversible.AI alignment ends may be contested.AI alignment value choices may remain implicit.Some technical alignment disagreements may conceal ethical disagreements.AI developers may face competitive pressure.Developers may see restraint as a competitive disadvantage.Coordination that could support restraint may be weak.Some actors may proceed despite recognising AI risks.Some AI uses may delegate human judgment.Some AI uses may reduce opportunities to practise judgment.Some AI uses may weaken human agency.AI’s possible moral standing may remain uncertain.Criteria for responding to possible AI moral standing may be absent.Potentially important AI moral considerations may receive inconsistent treatment.The still-unspecified AI development decision is defensible.The decision meets needs for timely benefits and security.The decision protects against unacceptable harm.Release the selected capability now.Defer release of the selected capability.Usable evidence about the decision is available.Decision makers have authority and incentives to act.Feasible alternatives are available.Make ethical aims and disagreements explicit.Explicit aims may expose assumptions hidden inside evaluations.Alternatives can be assessed against defensible criteria.Introduce a specific wisdom-derived practice with a mechanism and limits.The practice may change attention, motivation or judgment through its mechanism.A specified aspect of decision making may improve.Establish proportionate safeguards or coordination.Safeguards or coordination may change actors’ options or motivation.Restraint or redirection becomes practicable.Formal review may create unwarranted reassurance.Unwarranted reassurance may contribute to premature action.Require substantive criteria; completing review is not enough.Apply safeguards before actions whose harms could be irreversible.One decision practice can be responsibly tested in a defined setting.“Wise” remains underspecified.Evaluation criteria, disagreements and unacceptable outcomes are explicit.The causal diagnosis is unsupported.Evidence distinguishes the leading explanation from credible alternatives.The contribution from a tradition is vague.A specific practice, mechanism and limits are identified.The proposed change exceeds participants’ powers.A feasible intervention and relevant decision authority are identified.Success could be confused with completing the process.Outcome measures, guardrails and reconsideration criteria are defined.AI choices are discussed generally.Select one actor, decision and time horizon.Alternatives and responsibilities become concrete.Important premises remain implicit.Record aims, empirical claims and disagreements.The decision’s justification becomes inspectable.Several explanations could fit.Gather evidence that distinguishes the explanations.The intervention can target a better-supported mechanism.A bounded intervention is sufficiently justified.Specify a guarded test; conduct it only where appropriate.A responsible test can make evidence about effects available.Results are available.Compare results with the prediction and alternatives.The affected claims and next decision can be revised.and 38 connections

First shared-component edition
25 October 2026 · 2026-10-25.1

25Oct 2026

Four source-question hierarchies and three illustrative reasoning trees, using the established demonstration components. All reasoning remains proposed; no trial or outcome is recorded.

Added
People make informed, revisable choices about developing and using AI.Decision makers understand relevant benefits, harms and uncertainties.Decision makers can revise choices as evidence changes.Uncertainty about AI’s effects can make development and adoption choices difficult to justify.Some AI choices may proceed without an adequate account of their consequences.Decision makers use explicit review points to revisit AI choices against emerging evidence.AI choices can be revised when evidence changes their justification.and 4 connections

Original published statements, connections and attribution

{
  "projection_version": 4,
  "release_id": "2026-10-25.1",
  "snapshot_date": "2026-10-25",
  "title": "Wisdom and AI",
  "people": [
    {
      "name": "David Joseph",
      "role": "Requested this demonstration; has not endorsed the proposed reasoning"
    }
  ],
  "trees": {
    "goal": {
      "nodes": [
        {
          "id": "g-wise-ai",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "goal",
          "text": "People make informed, revisable choices about developing and using AI.",
          "label": "People make informed, revisable AI choices.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "How should we integrate AI into our lives and societies?",
              "basis": "inference"
            }
          ]
        },
        {
          "id": "g-understand",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "necessary_condition",
          "text": "Decision makers understand relevant benefits, harms and uncertainties.",
          "label": "Decision makers understand benefits, harms and uncertainties.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "How should we integrate AI into our lives and societies?",
              "basis": "inference"
            }
          ]
        },
        {
          "id": "g-revise",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "necessary_condition",
          "text": "Decision makers can revise choices as evidence changes.",
          "label": "Decision makers can revise choices as evidence changes.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        }
      ],
      "links": [
        {
          "id": "g-understand-necessary",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "from": [
            "g-understand"
          ],
          "to": "g-wise-ai",
          "relation": "necessary_for",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        },
        {
          "id": "g-revise-necessary",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "from": [
            "g-revise"
          ],
          "to": "g-wise-ai",
          "relation": "necessary_for",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        }
      ],
      "backlog": [
        {
          "ref": "g-wise-ai",
          "work": "review_goal",
          "label": "Review the goal and whether both conditions are necessary."
        },
        {
          "ref": "g-understand",
          "work": "investigate_wisdom",
          "label": "Investigate what wisdom traditions contribute.",
          "note": "Preserve disagreements between traditions and limits of analogy; the page asks rather than settles this."
        }
      ]
    },
    "current_reality": {
      "nodes": [
        {
          "id": "r-uncertainty",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "cause",
          "text": "Uncertainty about AI’s effects can make development and adoption choices difficult to justify.",
          "label": "Uncertainty can make AI choices difficult to justify.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        },
        {
          "id": "r-decisions",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "undesirable_effect",
          "text": "Some AI choices may proceed without an adequate account of their consequences.",
          "label": "Some AI choices may lack an adequate account of consequences.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        }
      ],
      "links": [
        {
          "id": "r-uncertainty-decisions",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "from": [
            "r-uncertainty"
          ],
          "to": "r-decisions",
          "relation": "contributes_to",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        }
      ],
      "backlog": [
        {
          "ref": "r-uncertainty-decisions",
          "work": "test_causality",
          "label": "Test whether uncertainty contributes to poorly justified choices.",
          "note": "Uncertainty could also prompt restraint or better deliberation; neither effect is established here."
        }
      ]
    },
    "future_reality": {
      "nodes": [
        {
          "id": "f-review",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "injection",
          "text": "Decision makers use explicit review points to revisit AI choices against emerging evidence.",
          "label": "Decision makers revisit AI choices at explicit review points.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        },
        {
          "id": "f-revision",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "kind": "desirable_effect",
          "text": "AI choices can be revised when evidence changes their justification.",
          "label": "AI choices can be revised when evidence changes their justification.",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        }
      ],
      "links": [
        {
          "id": "f-review-revision",
          "status": "proposed",
          "accepted_by": [],
          "contributors": [
            {
              "name": "AI assistant",
              "role": "assistant"
            }
          ],
          "from": [
            "f-review"
          ],
          "to": "f-revision",
          "relation": "causes",
          "source_refs": [
            {
              "source_id": "wisdom-ai",
              "locator": "Should we change pace or trajectory, and can we?",
              "basis": "inference"
            }
          ]
        }
      ],
      "backlog": [
        {
          "ref": "f-review-revision",
          "work": "test_causality",
          "label": "Test whether review points enable consequential revision.",
          "note": "Review points alone may not suffice: decision makers need authority, usable evidence and feasible alternatives. No benefit has been observed."
        },
        {
          "ref": "f-review",
          "work": "investigate_wisdom",
          "label": "Investigate how wisdom traditions might shape review practice."
        }
      ]
    }
  },
  "feedback": [],
  "history": [
    {
      "release_id": "2026-10-25.1",
      "date": "2026-10-25",
      "title": "First shared-component edition",
      "summary": "Four source-question hierarchies and three illustrative reasoning trees, using the established demonstration components. All reasoning remains proposed; no trial or outcome is recorded.",
      "changes": [
        {
          "type": "added",
          "refs": [
            "g-wise-ai",
            "g-understand",
            "g-revise",
            "g-understand-necessary",
            "g-revise-necessary",
            "r-uncertainty",
            "r-decisions",
            "r-uncertainty-decisions",
            "f-review",
            "f-revision",
            "f-review-revision"
          ]
        }
      ]
    }
  ],
  "publication_kind": "source_reconstruction",
  "authors": {
    "rufus": {
      "name": "Rufus Pollock",
      "role": "Leads the Wisdom and AI research project at Life Itself; posed the source research questions and has not reviewed or endorsed this proposed reasoning"
    }
  },
  "sources": [
    {
      "id": "wisdom-ai",
      "kind": "web_page",
      "author": "rufus",
      "url": "https://ai.lifeitself.org/",
      "title": "Wisdom and AI",
      "description": "A Second Renaissance research agenda led by Rufus Pollock of Life Itself. One governing question, four core questions and their subquestions explore alignment, risk and pace, broader impacts and consciousness. Wisdom traditions means especially Buddhism in this source, while inviting others. It asks questions rather than establishing answers. Captured 25 October 2026."
    }
  ],
  "views": {
    "goal": [
      {
        "id": "goal-example",
        "title": "An informed, revisable choice",
        "summary": "A suggested goal and two conditions to review. These are AI interpretations, not conclusions from the source.",
        "nodes": [
          "g-wise-ai",
          "g-understand",
          "g-revise"
        ]
      }
    ],
    "current_reality": [
      {
        "id": "current-reality-example",
        "title": "A possible difficulty",
        "summary": "A causal hypothesis that needs evidence; uncertainty can also prompt restraint or better deliberation.",
        "nodes": [
          "r-uncertainty",
          "r-decisions"
        ]
      }
    ],
    "future_reality": [
      {
        "id": "future-reality-example",
        "title": "A practice worth testing",
        "summary": "Review points might enable revision, but authority, usable evidence and feasible alternatives may also be needed.",
        "nodes": [
          "f-review",
          "f-revision"
        ]
      }
    ]
  },
  "overview": [
    {
      "type": "cards",
      "eyebrow": "A goal to review",
      "title": "Make informed choices, and keep them revisable.",
      "text": "The research agenda asks how we should develop and use AI wisely. This example proposes an operational goal and two possible necessary conditions.",
      "ids": [
        "g-wise-ai",
        "g-understand",
        "g-revise"
      ],
      "link": [
        "goal.html",
        "Inspect the proposed goal →"
      ]
    },
    {
      "type": "cards",
      "eyebrow": "A hypothesis to test",
      "title": "Can uncertainty leave choices poorly justified?",
      "text": "This is one possible problem to investigate, not a finding established by the source. Uncertainty could instead prompt restraint or better deliberation.",
      "ids": [
        "r-uncertainty",
        "r-decisions"
      ],
      "link": [
        "in-the-way.html",
        "Examine the causal hypothesis →"
      ]
    },
    {
      "type": "cards",
      "eyebrow": "A proposed change",
      "title": "Revisit decisions as evidence changes.",
      "text": "Explicit review points could make choices revisable. The link is still a hypothesis, and authority and feasible alternatives may be needed.",
      "ids": [
        "f-review",
        "f-revision"
      ],
      "link": [
        "change.html",
        "Inspect the suggested practice →"
      ]
    },
    {
      "type": "credits",
      "eyebrow": "Source questions and interpretation",
      "title": "Questions by the research project; proposed answers by the AI.",
      "text": "The source’s four research strands are preserved under Sources and limits. Every statement and connection in this small example is an AI inference awaiting review, not source-author endorsement."
    }
  ],
  "retired": []
}