Figuora

Question

Which human capabilities should remain uniquely human?

Not everything that can be automated should be. An investigation into which capacities we may want to protect as human by design.

OpenOpen QuestionUpdated August 5, 2026Human CapabilityFrontier TechnologyCivilization

Why it matters

Without a deliberate answer, market convenience will decide which human roles dissolve—including roles tied to legitimacy, care, and responsibility.

Current understanding

There is no consensus list of 'uniquely human' must-keeps. Strong candidates involve moral responsibility, intimate care, political authorization, and identity formation—but each can be mimicked or assisted by machines.

Known evidence

  • Public trust often collapses when high-stakes decisions feel unaccountable to a human locus of responsibility.
  • Care work involves affective and embodied dimensions poorly captured by task automation metrics.
  • Legal systems still require responsible agents even when tools advise outcomes.

Competing explanations

  • Nothing should remain uniquely human if machines outperform on welfare metrics.
  • Certain authority-bearing acts (judgment, pardon, declaration of war) must remain human for legitimacy.
  • Uniqueness is temporary; protect transitions and dignity, not eternal essences.
  • The real issue is ownership and oversight, not performing the act itself.

Figuora hypothesis

Hypothesis: capabilities that bind responsibility to a mortal, accountable person—final moral judgment, political authorization, and intimate care commitments—should remain human-led even when machines assist the surrounding work.

Labeled as a hypothesis—not established conclusion. It should be pressure-tested against counterarguments and unknowns below.

Counterarguments

  • Human judgment is biased and inconsistent; machine consistency may be more just.
  • Drawing bright lines may lag reality as hybrids become normal.
  • 'Mortal accountability' is romantic if institutions already diffuse responsibility.

Unknowns

  • How much assistance is compatible with meaningful responsibility.
  • Whether future humans will renegotiate dignity away from work entirely.
#ethics#automation#human-role

Sources

  1. 1. The Human Condition — Hannah Arendt (1958) · Action, work, and labor distinctions; philosophical frame.
  2. 2. What Computers Still Can't Do — Hubert Dreyfus (1972) · Classic limits-of-AI argument; partially overtaken, still generative.
  3. 3. Life 3.0 — Max Tegmark (2017) · Scenario menu for human-AI futures.