HIE LabHuman Intelligence Engineering
中文

Knowledge units

Claims

Clear, testable statements connected to the evidence that supports, qualifies, or challenges them.

Current status

AI evidence graph active

Initial focus

  1. 01

    Atomic and testable statements

  2. 02

    Evidence strength and limitations

  3. 03

    Revision history and uncertainty

Evidence graph

Published claims

AI-normalized propositions connected to the papers that support, qualify, or contextualize them.

  1. 01Contested

    Abstract reviewed · No human review

    Human-AI collaboration does not reliably outperform the better standalone human or AI; results depend strongly on the task and interaction design.

    A broad synthesis finds no universal synergy, while field and laboratory studies show meaningful gains in bounded tasks. The defensible conclusion is conditional, not categorical.

    AI assessment confidence90

    Evidence relationships

    • Qualifies
      When combinations of humans and AI are useful: A systematic review and meta-analysis

      Across 106 experiments, combinations were worse than the better standalone agent on average, with task type explaining important variation.

      Scientific evidence strength
      95
      AI relationship confidence
      95
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Supports
      Generative AI at Work

      AI assistance improved productivity in a specific customer-support setting, with heterogeneous effects by worker experience.

      Scientific evidence strength
      90
      AI relationship confidence
      91
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Supports
      Experimental evidence on the productivity effects of generative artificial intelligence

      Randomized access to ChatGPT improved speed and quality on bounded professional writing tasks, especially for lower-skill participants.

      Scientific evidence strength
      91
      AI relationship confidence
      93
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Qualifies
      Does the Whole Exceed its Parts? The Effect of AI Explanations on Complementary Team Performance

      Explanations increased acceptance of AI advice but did not increase complementary team performance.

      Scientific evidence strength
      87
      AI relationship confidence
      88
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
  2. 02Supported

    Abstract reviewed · No human review

    Expected access to external digital information changes memory strategy, reducing memory for content while preserving where it can be found.

    Controlled experiments and a broader cognitive-offloading framework converge on a shift from internal storage toward external-location memory.

    AI assessment confidence91

    Evidence relationships

    • Supports
      Google Effects on Memory: Cognitive Consequences of Having Information at Our Fingertips

      Participants remembered less content when future access was expected and remembered its location better.

      Scientific evidence strength
      92
      AI relationship confidence
      93
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Contextualizes
      Cognitive Offloading

      The cognitive-offloading framework explains why people move processing or storage demands into reliable external aids.

      Scientific evidence strength
      88
      AI relationship confidence
      91
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
  3. 03Emerging

    Abstract reviewed · No human review

    Smartphone presence or use has a small negative average association with cognitive performance, but the effect varies substantially by domain and context.

    Influential experiments report attention costs, but later synthesis finds only a small average effect with considerable heterogeneity.

    AI assessment confidence86

    Evidence relationships

    • Supports
      Brain Drain: The Mere Presence of One’s Own Smartphone Reduces Available Cognitive Capacity

      Two experiments found lower cognitive-task performance when participants phones were visible rather than in another room.

      Scientific evidence strength
      83
      AI relationship confidence
      84
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Qualifies
      Does the Brain Drain Effect Really Exist? A Meta-Analysis

      The pooled effect was small and heterogeneous, so a universal brain-drain interpretation is not supported.

      Scientific evidence strength
      88
      AI relationship confidence
      90
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Contextualizes
      The “online brain”: how the Internet may be changing our cognition

      The broader review places device effects within activity, population, and life-stage differences.

      Scientific evidence strength
      76
      AI relationship confidence
      85
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
  4. 04Emerging

    Abstract reviewed · No human review

    Online search can inflate perceived internal understanding by blurring the boundary between accessible external information and knowledge held in memory.

    Repeated experiments identify a metacognitive calibration effect, while broader reviews show that digital cognition outcomes remain context-sensitive.

    AI assessment confidence88

    Evidence relationships

    • Supports
      Searching for explanations: How the Internet inflates estimates of internal knowledge

      Across nine experiments, searching online increased estimates of personal explanatory knowledge, including for unrelated topics.

      Scientific evidence strength
      90
      AI relationship confidence
      91
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Contextualizes
      Cognitive Offloading

      Offloading decisions depend partly on metacognitive beliefs about internal ability and external reliability.

      Scientific evidence strength
      80
      AI relationship confidence
      88
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract
    • Contextualizes
      The “online brain”: how the Internet may be changing our cognition

      The review situates metacognitive effects within broader changes to attention, memory, and social cognition.

      Scientific evidence strength
      74
      AI relationship confidence
      84
      Source / review level
      Abstract reviewed
      Provenance
      codex-research-curation · initial-curation-v1
      Human review
      No human review
      Exact source scope
      title and abstract