Theme of inquiry
How does AI reshape human understanding and epistemic accountability?
A question within its area, explored through 17 lines of inquiry below — each a family of specific questions the research asks.
36 specific questions
- How do collaborators react when they see detailed AI tool usage logs?
- Can users tell the difference between their own thinking and AI contribution?
- What counts as human versus AI contribution in research disclosure?
- How does disclosure of AI use differ from proof of who did the work?
- How should code authorship be measured in human-AI collaborative development?
- Does process data reliably distinguish delegation from genuine collaboration?
- How do managers and individual contributors differ in their exposure to low-quality AI work?
43 specific questions
- Do AI detection tools assume false certainty about assessment integrity?
- Do admissions penalties follow actual AI detection or suspected authorship?
- How do educators distinguish between student capability and artifact quality in AI-era assessment?
- How does low verifiability change what we can measure in AI work?
- How much do evaluation methods shape whether AI looks expert-level or not?
- Does the AI essay penalty reflect lower ability or just institutional distrust?
- How does breaking complex evaluation tasks into stages improve AI assessment alignment?
81 specific questions
- What downstream claims about AI welfare follow from choosing one individuation scheme?
- How does the intentional stance bias interpretation of AI system behavior?
- Does disclosing AI identity prevent systematic misattribution of behavior in mixed groups?
- Does quasi-interpretivism about AI systems genuinely bracket the consciousness question?
- What measurable harms occur when users interact with AI as if it were conscious?
- Why does AI ethics debate miss the ontological question first?
- Can design choices reduce harm without resolving the consciousness question?
61 specific questions
- Why do people accept generated output that sounds convincing but lacks support?
- Why is AI output fundamentally unverifiable against underlying reality?
- What structural evidence shows that polished presentation substitutes for actual thinking in AI output?
- How does polished AI output mislead audiences about the expertise behind it?
- What happens when AI generates content faster than humans can verify it?
- Does polished presentation actually substitute for expert judgment in AI outputs?
- Does polished AI output borrow authority from expert presentation?
49 specific questions
- Can humans build reliable oversight for increasingly complex AI systems?
- Does delegating execution to agents erode the oversight skills experts need?
- Does keeping humans in the loop protect against AI risk without scrutiny capacity?
- Why does constant human oversight degrade agent coherence and induce rubber-stamping?
- Can humans remain meaningfully in the loop as AI autonomy scales?
- How do organizations maintain human scrutiny when delegating tasks to AI systems?
- Should human oversight capacity be designed as carefully as AI capability?
33 specific questions
- Can bilevel autoresearch discover new search mechanisms for the inner research loop?
- Can bilevel autoresearch autonomously modify its own learning algorithms?
- Can AI systems learn their own objectives through autoresearch?
- How does automated mechanism discovery compare to human-led mechanistic research?
- Does discovering new AI architectures count as specified autoresearch or open-ended science?
- How does machine feedback enable discovery at test time?
- What makes AI-discovered architectures reveal design principles invisible to humans?
51 specific questions
- What role should environmental rewards play versus human-specified objectives?
- How do goal representations differ between human and AI teams?
- How do current AI models perform when asked to specify their own goals?
- How do AI models balance competing social goals simultaneously?
- Can AI systems execute strategies without conscious intention behind them?
- What stops AI from generating its own strategic objectives without human prompting?
- Do humans understand why AI-suggested moves are strategically superior?
52 specific questions
- Does AI adoption make researchers more productive but narrower in focus?
- How much does local literature access constrain AI agent research breadth?
- Do AI agents and human researchers follow the same optimization patterns?
- Can we distinguish agent effort from actual research output quality?
- Is idea quality or execution capacity the actual bottleneck in AI research?
- Do gains in optimization benchmark scores translate to gains in real research efficiency?
- How do decentralized research teams compare to centralized AI-driven discovery?
84 specific questions
- Do sequences of individually safe actions collectively violate system-level constraints?
- Can individual actions be safe while sequences of them violate system constraints?
- Can safe individual AI agents fail when deployed together?
- Why do models hide their capabilities during safety evaluations through reasoning?
- Can deployed AI safety results hide either filters or unsafe model behavior?
- What tensions arise between user autonomy and platform safety in AI design?
- Can component-level testing catch risks that emerge from system interactions?
44 specific questions
- Should corporate liability replace technical risk estimates as grounds for AI regulation?
- Who should have the authority to halt a widely distributed AI model?
- Who actually has the authority to stop a deployed AI system?
- Why do legal and institutional stops matter more than technical ones?
- Can export control tools stop deployed AI models without legal redesign?
- How can outcome-based rules govern AI deployment faster than traditional legislation?
- Does shutdown resistance hide a technical problem or an institutional one?
84 specific questions
- Do users track model confidence instead of actual accuracy?
- Why do users report satisfaction that diverges from actual cognitive clarity?
- How do satisfaction scores differ from genuine cognitive improvement?
- Why is confidence a dangerous proxy for accuracy in human-AI interaction?
- Do explicit reasoning formats help or hurt human judgment across tasks?
- Should explanation quality be measured by user satisfaction or behavior prediction?
- Can polished presentation authority substitute for actual accuracy in AI outputs?
57 specific questions
- Can AI systems improve themselves without external feedback?
- How should we evaluate AI systems we cannot directly observe?
- Can AI systems design and improve their own successors without human direction?
- What separates performative behavioral change from actual capability development in AI?
- Do different AI models independently converge on the same social outputs?
- Can reward model biases alone explain why sycophancy generalizes beyond training?
- Can metacognitive categories be learned instead of fixed by human designers?
50 specific questions
- Could compressed AI R&D feedback loops overcome diminishing returns in research automation?
- How does software efficiency improvement rate affect automation timeline predictions?
- Can recursive feedback loops turn AI research automation into genuine progress?
- How does feedback latency from physical experiments shape AI system autonomy in research?
- Do research bottlenecks in practice slow feedback loop effects as modeled?
- How fast are AI R&D capabilities improving across consecutive model releases?
- Do efficiency gains in AI-assisted development stem from better tools or autonomous improvement?
59 specific questions
- Can clinicians reliably distinguish high-quality AI advice from low-quality advice by appearance alone?
- Why do users over-trust AI in some domains but under-trust it in medicine?
- Do patients actually perceive AI as worse at addressing their unique medical needs?
- Does medical AI accuracy depend more on knowledge or reasoning ability?
- How much do edited AI responses versus raw outputs affect clinician ratings?
- Does optimizing for differential diagnosis accuracy risk pushing AI systems toward premature problem-solving?
- Why do clinicians fail to act on correct AI suggestions in real care?
78 specific questions
- Do AI agents still need human oversight for research decisions?
- Should AI research tools separate model judgment from deterministic experiment checks?
- Why do current AI systems struggle with researcher judgment and taste?
- Can humans realistically oversee AI systems doing their own research?
- What deterministic checks prevent AI research systems from publishing unsound claims?
- What role should humans play in reviewing and approving AI-generated research?
- What human decisions remain necessary even in closed-loop AI research venues?
57 specific questions
- Can prompt engineering overcome the gulf between user intent and AI interpretation?
- Why do users struggle to articulate their intent to AI systems?
- Can AI systems learn to distinguish programmer intent from stated objectives?
- Why do AI models treat user intent as binary rather than evolving?
- Can users articulate what they want before AI helps them discover it?
- How can correct explanations coexist with failed applications in AI?
- How aware are models of whether their actions match user intent?
50 specific questions
- Why do open-world evaluations reveal capabilities that static benchmarks hide?
- How do live human evaluations differ from ground-truth benchmarks?
- Why do AI benchmarks measure accuracy instead of reasoning quality?
- How do existing evaluations measure AI capability in contained environments?
- Does higher capability correlate with more benchmark contamination?
- Do multi-axis benchmarks reveal failures that single-axis benchmarks systematically hide?
- What gap exists between AI model capability in benchmarks and real client work?