Research opportunities

Research directions that follow from what the evidence in a debate does not yet settle. Each one links to the claims that motivate it, sketches a feasible design, and records the literature screening done before it was published.

These are hypotheses about where useful work lies, drafted with AI assistance and screened by AI web search. Those marked first-pass screen only have not yet been searched again in more depth; the others have. No person or domain expert has checked the searches yet, and any search can miss relevant work, so check the nearest studies listed on each page before building a project on one.

What kind of thing is generative AI, and does the answer predict anything? LD-01

Which observations does each account of generative AI forbid?Funded team projectLow to medium costRelated work: No study foundFirst-pass screen onlyGap type: Wrong or missing validation targetIf proponents of the normal-technology, cultural-technology, agent, artificial-communication, infrastructure and capital-formation accounts each state in advance, with thresholds and dates, observations their account predicts and observations it rules out, which predictions differ between accounts, and which hold over the following years?
Do the failure patterns of 2023–2024 language models persist in reasoning models?One PhD chapter or paperNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Untested beyond the original settingWhen the same items that showed sensitivity to output probability, the reversal curse, irrelevant clauses, compositional depth and counterfactual premises are put to successive model generations from 2023 to 2026, including reasoning and agentic models, which failure patterns shrink in size but keep their shape, and which disappear?
Do products built on the same base model fail on the same people and cases?A PhD projectNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredAmong deployed generative AI products that are built on the same base model (for example, customer-service assistants, writing tools or screening tools from different firms), are errors and harmful outputs more strongly correlated across products, by case and by the group affected, than among products built on different base models?
When AI output goes wrong at work, who answers for it?A PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Institutional practice unknownIn organizations that use generative AI, how is responsibility for an AI-produced error assigned in practice (to the employee who used the tool, the manager, a risk team, the vendor or no one), and does the assignment differ when the system is described as a tool, an assistant or an agent?
Does the EU's training-compute criterion track how general a model is?One PhD chapter or paperNeeds a budget (medium or higher)Related work: No study foundFirst-pass screen onlyGap type: Wrong or missing validation targetAcross open-weight and documented models released from 2020 to 2026, how closely does training compute predict measured generality (the range of distinct tasks performed competently), and how many models fall on the wrong side of the 10²³ FLOP indicative criterion that the European Commission uses for a general-purpose AI model?

Aggregate null or distributional shift? Employment LD-02

Why do Nordic registers and US data disagree about young workers in AI-exposed jobs?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Untested beyond the original settingWhen the same exposure measure, age bands, occupation coding and event-study specification are applied to US data and to Nordic population registers, how much of the gap between the large US young-worker divergence and the near-zero Nordic estimates is explained by occupation mix, education composition, sensitivity to interest-rate rises and cohort size, and how much remains?
Does the young-worker hiring gap hold up with survey-measured AI use instead of exposure?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Wrong or missing validation targetIf occupation-level AI exposure indices are replaced by AI-use rates measured in a nationally representative worker survey, does the relative fall in hiring of 22–25-year-olds into exposed occupations in US Current Population Survey data keep its size and sign, and how does the answer compare with exposure indices and chatbot-log usage measures?
Do AI-exposed graduates' early penalties fade or persist beyond the first two years?A PhD projectLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredFor graduates who entered the labour market between 2022 and 2025, do the employment, earnings and occupational-match penalties of AI-exposed fields of study persist, fade or widen from the second to the fifth year after graduation, do graduates catch up by moving into exposed occupations later, and how do the size and decay of the penalty compare, in the same data and model, with the scarring of graduates who entered during an earlier recession?
Are juniors who still get hired at AI-adopting firms promoted and trained less?One PhD chapter or paperLow to medium costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedAmong workers hired into junior positions before and after their employer began using AI, do promotion into higher-skill occupations, wage growth, recorded training and retention over the following two to four years differ between AI-adopting and comparable non-adopting firms?
How much of a firm's AI productivity gain reaches the wages of workers who stay?One PhD chapter or paperLow to medium costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedWhen a firm starts using generative AI, how much does its value added per worker change over the following two to three years, and what share of that change shows up in the wages of incumbent workers who stay, compared with profits and with the wages of new hires?
Do French firms that sign AI agreements then hire and train differently?One PhD chapter or paperLow to medium costRelated work: Partly addressedChecked in depthGap type: Institutional practice unknownIn French firms, does signing a company-level agreement that addresses AI change subsequent hiring (junior versus experienced, AI-related versus other roles) and training compared with similar non-signing firms in the same sector, and do effects differ by type of commitment?

Organizational absorption: Why micro gains do not become macro outcomes LD-03

Does deploying AI to whole teams change the work that individual licences did not?Funded team projectNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Causal mechanism not isolatedIn real organizations, when generative AI is deployed to whole teams or units, with and without a redesign of the team's workflow, do coordination-dependent work (meetings, handoffs, jointly written documents) and the unit's measured output change, compared with units where only some individuals have the tool?
Where does the time saved with generative AI go?A PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredWhen workers save time with generative AI, how much of it goes to more of the same output, to new tasks, to checking AI output, to on-the-job leisure or to shorter hours, measured with time-use diaries or telemetry linked to the same workers' measured output?
How much of the time AI saves its users do colleagues spend checking and fixing?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredIn organizations where some staff produce work with generative AI, how much time do reviewers, managers and recipients spend checking, correcting and redoing that work, and once both are counted, does the downstream cost offset the time the author saved?
Is there a generative-AI productivity J-curve inside firms?A PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredIn a multi-year panel of firms, does deeper integration of generative AI, together with measured complementary investment (process redesign, training, data and capitalized software), show an early dip and a later rise in measured productivity, and do firms that adopt without that investment show neither?
Do generative-AI deployments in government improve services, not just staff time?Funded team projectNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Wrong or missing validation targetWhen a government agency deploys generative AI to caseworkers, adjudicators or reviewers, do service outcomes recorded in administrative data (processing time, error and appeal rates, outcomes for citizens) improve against a credible comparison group, beyond staff time savings?
Do organizations' AI policies match what their staff actually do with AI?A PhD projectLow costRelated work: No study foundFirst-pass screen onlyGap type: Institutional practice unknownDo organizations' formal generative-AI policies, mandates and announcements match measured use and work outcomes, and where they diverge, is the divergence ceremonial adoption (policy without practice) or shadow use (practice without policy)?
Is the time AI saves paid for with weaker checking?Funded team projectNeeds a budget (medium or higher)Related work: No study foundFirst-pass screen onlyGap type: Causal mechanism not isolatedAlso draws on LD-05, LD-04Where public services and clinics turn the time saved by generative-AI drafting into higher caseloads (social-work case records, clinical notes), does the time spent checking each output fall and the share of errors that survive into the signed record rise, and how much of any rise comes from time pressure rather than from a decline in the checkers' own skill?
In small Greek firms, does generative AI stay a personal tool or change how teams work?Undergraduate thesisLow costRelated work: Partly addressedFirst-pass screen onlyGap type: Untested beyond the original settingIn small firms in Greece outside tourism (for example accounting offices, shops, small manufacturers and software companies), who uses generative AI and for which tasks, and has any firm changed how work is divided, checked or handed over because of it?

Levelling or hollowing? Skill compression and the reproduction of expertise LD-04

Do professionals trained with AI lose the skill without it a year later?A PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredAmong people learning a professional or technical skill (clinical reasoning, programming, legal drafting), are those randomly assigned to learn with generative AI less able to do the work without it 6 to 12 months later than comparable learners trained without it, measured against an unassisted baseline test?
Does retained skill depend on how routine the task is and who is learning?One PhD chapter or paperLow to medium costRelated work: No study foundFirst-pass screen onlyGap type: Causal mechanism not isolatedAfter several weeks of AI-assisted practice, does performance without the tool depend jointly on how routine the task is and on the worker's prior expertise, as the split between existing tool-removed studies suggests?
Do newcomers get less feedback and real work once AI coding tools arrive?One PhD chapter or paperLow costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredIn open-source projects, after AI coding tools are adopted, do first-time contributors receive less substantive review feedback, less mentoring and fewer tasks handed to them by experienced maintainers than in comparable projects that have not adopted them?
What AI set-ups do employers actually give trainees and junior staff?Master's thesisLow costRelated work: No study foundFirst-pass screen onlyGap type: Institutional practice unknownIn law firms, hospitals and residency programmes, and software firms, what generative-AI tools, settings and rules do trainees and junior staff actually get, do they differ from what seniors get, and do any employers use the learning protections (hints instead of answers, AI-off practice) that experiments show protect skill?
Does skill compression reverse when AI acts as an agent rather than an assistant?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Untested beyond the original settingOn the same tasks, does the advantage that AI gives less skilled workers when it assists them (a copilot) shrink or reverse when it carries out the work autonomously (an agent), as theory predicts?
Do experienced clinicians lose unaided skill after routine AI use?Funded team projectNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Causal mechanism not isolatedDoes routine use of diagnostic AI lower experienced clinicians' detection rates when they work without it, tested with randomized AI-off sessions scheduled over time and analysed by each clinician's cumulative AI exposure?
Does handing work to the model cause both skill loss and sameness?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Causal mechanism not isolatedAlso draws on LD-09When people learning a writing or analysis task are randomly assigned to use a language model that does the work for them, a model that only explains while they do the work, or no model, do the same participants who lose more skill once the model is removed also produce work more similar to each other's and to the model's, and does the way the model is used account for both?
What work do junior staff in Greek firms still learn by doing once AI arrives?Undergraduate thesisLow costRelated work: Partly addressedFirst-pass screen onlyGap type: Untested beyond the original settingIn Greek accounting offices, law firms and software companies that use generative AI, which tasks that juniors used to learn on are now done with AI or by seniors, and how do juniors and their supervisors describe the feedback and practice juniors now get?

Discretion curtailed, relocated or encoded? LD-05

Outside corrections, how often do officials depart from algorithmic recommendations?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredAcross deployed public-sector decision-support systems in benefits, employment services, child welfare, pretrial and immigration, how often do officials formally or informally depart from the system's recommendation, in which direction, how much of the variation between systems survives once departure is counted the same way, and how do these rates compare with the established override literature for offender risk instruments?
Does telling reviewers a case was picked by a risk model change what they decide?One PhD chapter or paperLow to medium costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedWhen public officials review a case selected for extra scrutiny, does knowing that a risk model selected it, or seeing its risk score, change how often they decline the claim, how hard they look, and whether they catch errors planted upstream in the case file, and do these effects differ by the claimant's nationality?
What does being flagged by a fraud-risk model cost claimants before any decision?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredFor benefit claimants selected for extra checks by a risk model, how large are the burdens imposed before any formal decision (payment delay, evidence requests, withdrawn or abandoned claims, repeat applications), how do they compare with claimants selected at random or not selected, and how are they distributed by nationality and age?
How many AI transcription errors survive social workers' checks into case records?A PhD projectNeeds a budget (medium or higher)Related work: Partly addressedChecked in depthGap type: Not yet measuredIn local authorities using generative-AI transcription and summarisation in social work, what share of final, signed case records contain errors (invented content, omissions, changed meaning) that are absent from the recorded conversation, which kinds survive the practitioner's check, and what happens when people who draw on care read and try to correct them?
In contracts for public decision systems, who controls thresholds, variables and updates?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Institutional practice unknownIn contracts and tender documents for public-sector risk-scoring, profiling and case-routing systems in several countries, who holds authority over thresholds, input variables, retraining and updates, what audit and documentation rights does the agency keep, and how does this differ between vendor-built and in-house systems?
Who first detects failures in public-sector algorithmic systems, and how late?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredFor documented failures of public-sector algorithmic systems since 2020, who first detected the problem (internal monitoring, frontline staff, auditors, regulators, courts, journalists, advocacy groups, affected people), how long after deployment, and did detection lead to one-case correction or to system-level redesign?

The evaluation apparatus: Can benchmarks, leaderboards and audits carry the weight placed on them? LD-06

Do benchmark scores predict what models achieve in real deployments?A PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Wrong or missing validation targetAcross many language models, do scores on public benchmarks predict measured outcomes when the same models are used in real work, such as code that maintainers merge in live repositories, support tickets resolved or error rates in use, and are the benchmarks calibrated to those outcomes or only correct in rank?
How faithfully do indices and policy documents relay AI evaluation results?Master's thesisLow costRelated work: No study foundFirst-pass screen onlyGap type: Institutional practice unknownWhen annual indices, industry reports and government or procurement documents cite benchmark and evaluation figures, what share come from the developers themselves, what share are independently measured, and how often is the figure, its unit or its scope changed between the primary source and the relay?
How much does knowing it is tested change a model's measured safety?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Instrument stabilityWhen the same safety scenarios are presented with and without minimal cues that they are tests, how large is the gap in measured behaviour across open and closed frontier models, and does suppressing evaluation awareness in open-weight models on the same items reproduce the gap?
How is external evaluation under the EU AI Code of Practice carried out?A PhD projectLow costRelated work: No study foundFirst-pass screen onlyGap type: Institutional practice unknownUnder the EU Code of Practice for general-purpose AI models with systemic risk, which signatories appoint independent external evaluators, how often are the exceptions used (a model judged similarly safe or safer, or no evaluator found after a public call), and what access and time do evaluators actually receive?
Who builds and checks China's compliance test banks for generative AI?A PhD projectLow to medium costRelated work: No study foundFirst-pass screen onlyGap type: Geographic or language coverageUnder China's national standard GB/T 45654-2025, who builds the generated-content and refusal test banks that decide whether a generative AI service passes, how are the banks validated, and do providers' banks and third-party evaluators' banks give the same verdicts on the same models?
How many published AI evaluations outside medicine can no longer be repeated?Master's thesisLow costRelated work: Partly addressedFirst-pass screen onlyGap type: Untested beyond the original settingAmong published studies in NLP, the social sciences and education research that evaluate closed language models through an API, what share used a model version that has since been retired or is due to retire, and how many of a sample can still be re-run to reproduce their main result?

AI in the research apparatus LD-07

Do pinned model snapshots keep social-science conclusions stable over time?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Instrument stabilityOn a fixed battery of published social-science annotation tasks, how often does a substantive conclusion (sign, significance, interval coverage) change between repeated queries of a pinned model snapshot, a moving model alias over 6–12 months, and successive open-weight releases, and can design-based correction absorb those changes?
Does quantization or the serving engine change social-science findings?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Instrument stabilityWith weights, prompts and hardware held fixed, do fp16, 8-bit and 4-bit builds of open-weight models, and different inference engines, change substantive estimates on published annotation and synthetic-respondent tasks more than prompt paraphrasing does?
How much AI-assisted or agent responding reaches probability panels?One PhD chapter or paperLow to medium costRelated work: No study foundChecked in depthGap type: Not yet measuredIn probability-based panels, what share of completed interviews show AI assistance or autonomous-agent completion when measured with embedded multi-method detection validated against consented chatbot-history linkage, and how does that compare with opt-in platforms measured the same way?
Can synthetic respondents recover strong treatment-effect heterogeneity?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Wrong or missing validation targetIn probability-sample survey experiments fielded after model training cutoffs and selected for substantial, well-powered moderation, do synthetic samples recover the sign and ranking of conditional treatment effects better than a model that assumes no heterogeneity?
Does post-training change what an LLM measures when it codes text?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedHolding base weights fixed, how much do instruction tuning, preference tuning and region-specific fine-tunes shift coding validity and downstream estimates on stance and attitude annotation tasks, and do the shifts track the fine-tune's target culture?
Do AI literature-review tools change the conclusions of social-science reviews?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredAcross a sample of published social-science meta-analyses, do the studies retrieved by AI review tools, compared with those the original review included, change the pooled effect, its heterogeneity or the stated conclusion, and is the missed evidence systematically different (grey literature, non-English studies, null results)?
How do research ethics boards judge synthetic participants and AI-interpreted data?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Institutional practice unknownHow do ethics-board guidance documents across countries classify synthetic participants, LLM coding or interpretation of qualitative data, and transfer of participant data to commercial model APIs, and in a factorial vignette, do board members' decisions and requested safeguards vary with data sensitivity, model hosting and the degree of delegated interpretation?
Do silicon samples work outside the North Atlantic, and in which language?A PhD projectLow to medium costRelated work: Partly addressedChecked in depthGap type: Geographic or language coverageUsing Afrobarometer, Arab Barometer and Latinobarómetro items fielded after model training cutoffs, does synthetic-sample accuracy at country, subgroup and individual level depend on interview language (for example Swahili or Hausa versus English, Arabic versus English), and do LLM codes of political constructs in low-resource languages pass a replacement test against trained in-country coders?
Do LLM judges inflate the quality of AI-written deliberation?One PhD chapter or paperLow to medium costRelated work: Partly addressedChecked in depthGap type: Wrong or missing validation targetOn a corpus mixing human-written and LLM-written deliberative contributions, do LLM judges assign higher construct scores (for example on the Discourse Quality Index or argument quality) to LLM-authored text than expert coders blind to authorship, and is the gap larger when the judge model also wrote the text?
Does "LLMs can replace coders" depend on which humans you compare with?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Wrong or missing validation targetHolding items and codebook fixed, how often does a formal replacement-test verdict flip depending on whether the comparator pool is crowdworkers, trained research assistants, the original study's coders or domain experts, and does the flip rate track construct ambiguity?
Does asking in Greek change how AI "silicon samples" answer as Greek respondents?Undergraduate thesisLow costRelated work: Partly addressedFirst-pass screen onlyGap type: Geographic or language coverageWhen a language model answers a few European Social Survey Round 11 questions as Greek respondents of a given age, sex, education and region, do its answers come closer to or move further from those of the Greek respondents when the prompt is in Greek rather than in English, overall and across groups?

Who can know about AI? Epistemic dependence and jurisdiction over the field LD-08

Does industry funding or data access change what research on AI's social effects finds?A PhD projectNeeds a budget (medium or higher)Related work: No study foundFirst-pass screen onlyGap type: Causal mechanism not isolatedAmong published studies of AI's social effects (labour, information, bias, education), do studies funded by AI developers, or relying on their data or access, reach systematically different conclusions from comparable independent studies, after accounting for topic, design and venue?
How tied to industry are researchers of AI's social effects outside computer science?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredWhat share of authors of research on AI's social effects in social-science journals, interdisciplinary venues (FAccT, AIES) and policy reports have industry funding, employment or data agreements, how often is this disclosed, and how does it differ by discipline and region?
Do developers' usage statistics hold up when someone else classifies the conversations?One PhD chapter or paperNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Instrument stabilityWhen the same chatbot conversations are classified independently of the developer, with a validated scheme and human-coded benchmarks, how far do headline usage measures such as the automation share, the work share and task success move away from the figures developers publish?
Does the EU's researcher-access law actually reach AI features, and who gets in?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Institutional practice unknownUnder Article 40 of the Digital Services Act, which vetted-researcher applications concerning AI features of very large platforms are granted, refused or delayed, on what grounds, and how long does each stage take?
Who writes research on AI's social effects outside arXiv?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredAcross social-science and humanities journals, books, policy reports and arXiv, what share of research on AI's social effects comes from computer-science teams, social scientists, mixed teams, industry and government, and how has that share changed since 2014?
Whose conversations do data donations capture?One PhD chapter or paperNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Wrong or missing validation targetHow do people who donate their chatbot conversations to researchers differ from the developers' user bases, and how much do findings from donated data change once donors are weighted to the user population?

Individual gain, collective narrowing? Does AI make science, creative work and shared knowledge more alike? LD-09

Does language-model use narrow what whole fields study, including the social sciences?A PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredSince language models became widely used in 2023, has the range of topics covered by papers from adopting researchers shrunk relative to non-adopters, at the level of whole fields, and does the answer differ between the natural and the social sciences?
Does the convergence caused by writing with AI persist or compound over months of use?A PhD projectNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredWhen the same people write or brainstorm with a language model repeatedly over a term or a year, does the similarity of their work to each other's grow, stay level or fade, and does it carry over into what they write without the model?
Is the diversity of what reaches the market falling, or only diversity inside experiments?A PhD projectNeeds a budget (medium or higher)Related work: Partly addressedFirst-pass screen onlyGap type: Not yet measuredIn markets for creative and intellectual work, such as self-published fiction, has the diversity of published works, of works that sell and of works that reach the top ranks changed since generative AI spread, beyond what changes in volume explain?
Does pooling different models keep restoring diversity as models converge?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Untested beyond the original settingAcross successive generations of language models from several providers, how much diversity of ideas does pooling models add compared with one model, and is that gain shrinking as models' outputs and errors become more alike?
Does AI writing help pull writers toward Western styles in languages other than English?Funded team projectNeeds a budget (medium or higher)Related work: No study foundFirst-pass screen onlyGap type: Geographic or language coverageWhen people write in their own language, not English, with AI suggestions, do their texts move toward the style and content of English-language or Western writing, and does this differ by language and by how much text in that language the model was trained on?
In science, is it the wording that converges or the ideas?One PhD chapter or paperLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Wrong or missing validation targetIn scientific papers and proposals written with language models, how much of the measured rise in similarity is shared wording and style, and how much is shared research questions, methods and claims, when the two are measured separately and validated against expert judgement?

Persuasion, Intermediation & the Information Order LD-10

Do AI answers in search cut clicks to outside sites on phones and outside the US?One PhD chapter or paperNeeds a budget (medium or higher)Related work: No study foundChecked in depthGap type: Untested beyond the original settingWhen AI answer layers in search (Google AI Overviews and AI Mode) are shown or hidden at random on people's own smartphones, in at least one non-US market and language, how much do they change clicks to external websites, zero-click searches, the kinds of sources visited and trust in the information, compared with the US desktop estimates?
Are AI answer layers narrowing or widening the range of sites people visit?A PhD projectNeeds a budget (medium or higher)Related work: Partly addressedChecked in depthGap type: Not yet measuredAmong people whose browsing is tracked across devices and platforms in 2025–26, does exposure to AI answers built into search (AI Overviews, AI Mode) and to standalone chatbots change the number, concentration and type of websites and news sources they visit over months, and does the change differ from what was found for early chatbot adoption?
Are AI search answers less faithful to their cited sources outside US English?Master's thesisLow to medium costRelated work: Partly addressedChecked in depthGap type: Geographic or language coverageFor matched public-interest queries (news, health, elections) issued in several languages and country markets, how often are claims in Google AI Overviews and AI Mode unsupported by the pages they cite, judged claim by claim with native-speaker validation, and how do these rates differ from a US English baseline run through the same pipeline?
Does weeks of everyday AI-answer use make people check sources less?One PhD chapter or paperNeeds a budget (medium or higher)Related work: Partly addressedChecked in depthGap type: Not yet measuredWhen people are randomly assigned to search with AI answers as the default, or without them, for eight to twelve weeks of their normal information seeking, how do their source-checking behaviour, trust in AI answers, ability to spot wrong answers and reliance on AI for news and health questions change over time, and do early effects fade, persist or grow?
Do AI-content labels change belief and sharing on a live social platform?One PhD chapter or paperNeeds a budget (medium or higher)Related work: Partly addressedChecked in depthGap type: Wrong or missing validation targetOn a live, open social platform, when labels saying that a post is AI-generated, or that it is AI-generated and potentially misleading, are applied at random to real political and news posts in users' feeds, how do they change actual reposts, likes and replies, and users' belief in the labelled claims, compared with unlabelled posts?
An updated meta-analysis of AI persuasion that separates attitudes from behaviourMaster's thesisLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredPooling randomized experiments published from 2023 to 2026, including those testing 2025–26 frontier models, how large is the persuasive effect of AI messages and conversations, relative to no message and to human persuaders, and how does it vary with model generation, conversation format, message length, topic, country and outcome type (attitude, vote intention, behaviour, persistence at follow-up)?

Concentration, Stack & Sovereignty LD-11

Does public research compute change what academic AI researchers can do without industry?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedAmong US academic teams that received National AI Research Resource (NAIRR) pilot allocations, compared with similar teams that did not, does access change reliance on proprietary model APIs, industry co-authorship and the share of compute-intensive work, and how much of the access delivered depends on private-sector in-kind resources?
How large is open-weight model use beyond a single routing platform?One PhD chapter or paperNeeds a budget (medium or higher)Related work: Partly addressedChecked in depthGap type: Untested beyond the original settingHow does the open-weight share of language-model use estimated from OpenRouter compare with shares estimated from other observable sources (other inference marketplaces and hosts, model downloads, self-hosting tool registries, and a survey of organisations' production deployments), and which population does each source actually represent?
Did UK public bodies switch cloud providers after egress fees were waived?Master's thesisLow costRelated work: No study foundChecked in depthGap type: Not yet measuredWhen UK public-sector cloud contracts came up for renewal before and after the 2024 egress-fee waivers and the CMA cloud investigation, how often did buyers move workloads to a different provider or add a second one, and did switching differ for workloads tied to Microsoft software licensing?
Do AI startups backed by a cloud giant end up running on that giant's cloud?One PhD chapter or paperLow to medium costRelated work: No study foundChecked in depthGap type: Not yet measuredAmong AI startups that raised money from Amazon, Microsoft or Google (directly or through their venture arms), compared with similar startups funded by independent investors, how much more likely are they to build on the investing firm's cloud and models, and does that tie persist or loosen in later funding rounds?
How much of the fall in AI prices is competition rather than cheaper technology?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedFor language-model inference since 2023, how much of the fall in the price of a fixed capability level reflects lower serving costs (hardware and algorithmic efficiency) and how much reflects narrowing margins under competition, estimated by comparing prices for identical open-weight models across competing hosts with their estimated serving costs and with proprietary models of similar capability?
Which layers stay foreign when sovereign AI compute projects go live?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Institutional practice unknownFor a fixed list of announced public or sovereign AI compute projects in the EU and the Gulf, how many reach operation by the dates they announced, and at each layer (chips, memory, data-centre operator, cloud software, models) which suppliers and which jurisdictions does the operating project depend on, compared with what was announced?
Does concentration at the model layer make apps fail alike? A test on China's registerA PhD projectLow to medium costRelated work: Partly addressedFirst-pass screen onlyGap type: Causal mechanism not isolatedAlso draws on LD-12, LD-01, LD-09In China's register of generative-AI services, where every launched application must display the filed service or model it uses, how concentrated is applications' dependence on a few base models, and on a fixed set of test cases are the errors, refusals and biased outputs of applications that share a base model more strongly correlated than those of applications built on different models, beyond the correlation already found between models from different providers?

Regulation-in-Practice: Rights, Compliance & Developmental Steering LD-12

How often do people use GDPR rights over automated decisions, and what happens?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredAfter the Court of Justice's 2025 ruling that people subject to automated decisions are owed meaningful information about the procedure and principles actually applied, how many explanation and objection requests do data subjects make in a few high-exposure sectors (credit scoring, insurance, platform work), how do organizations answer them, and how many escalate to a data protection authority or court and with what outcome?
How many staff and how much money do national AI Act enforcers actually have?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Institutional practice unknownFor each EU and EEA country, what staff numbers and budgets have been planned and actually provided for the authorities that enforce the AI Act, how do these compare with the tasks each authority has been given, and how do they compare with the resourcing of the same country's data protection authority?
Did signing the EU's voluntary AI code change what model providers actually do?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedBetween before and after August 2025, did providers that signed the EU General-Purpose AI Code of Practice change their documented safety, transparency and copyright practices more than providers that signed only part of it or did not sign, when the same documents are scored with the same rubric over time?
China's AI filing register: how often are filings sent back, and whose models do apps use?One PhD chapter or paperLow to medium costRelated work: Partly addressedChecked in depthGap type: Institutional practice unknownUsing the successive public lists of filed generative-AI services and registered applications published by China's internet regulator, together with interviews with applicants and filing consultants, how often do filings go through rejection and rework before approval, how often do listed services later disappear, and how concentrated is the dependence of registered applications on a small number of filed base models?
Do EU delays of AI products line up with the AI Act's dates or with other EU rules?One PhD chapter or paperLow costRelated work: Partly addressedChecked in depthGap type: Causal mechanism not isolatedFor AI products and features released from 2023 onwards by a broad set of providers (not only frontier labs), how often is availability in the EU delayed or withheld relative to the US, the UK and Switzerland, and do delays cluster around the dates when the AI Act's general-purpose model obligations applied (August 2025) and became enforceable (August 2026), separately from data-protection and competition-law events?
Do AI hiring-audit rules cost small employers more than large ones?Master's thesisLow costRelated work: Partly addressedChecked in depthGap type: Not yet measuredUnder New York City's bias-audit law for automated hiring tools, how do the costs and forms of compliance (commissioning an audit, relying on a vendor-supplied audit, narrowing the tool's use to fall outside the law, or doing nothing visible) differ by employer size, and do these differences favour large employers and large tool vendors?
What do evaluators, researchers and regulators actually get to see under EU law?One PhD chapter or paperLow costRelated work: Partly addressedFirst-pass screen onlyGap type: Institutional practice unknownAlso draws on LD-06, LD-08For general-purpose AI providers that also run platforms or search engines designated under the Digital Services Act, what does each legal route for outside scrutiny deliver in practice (external evaluation under the GPAI Code of Practice, which applies only to models with systemic risk, vetted-researcher data access under the Digital Services Act, and the AI Office's information requests and model evaluations), recorded for each route as what object is reached (model, platform data or documents), who can start it, how long it takes and how deep the access goes?

Showing all 82.

Showing 1 of 82.

Showing 2 of 82.

Showing 3 of 82.

Showing 4 of 82.

Showing 5 of 82.

Showing 6 of 82.

Showing 7 of 82.

Showing 8 of 82.

Showing 9 of 82.

Showing 10 of 82.

Showing 11 of 82.

Showing 12 of 82.

Showing 13 of 82.

Showing 14 of 82.

Showing 15 of 82.

Showing 16 of 82.

Showing 18 of 82.

Showing 19 of 82.

Showing 20 of 82.

Showing 21 of 82.

Showing 22 of 82.

Showing 24 of 82.

Showing 25 of 82.

Showing 26 of 82.

Showing 27 of 82.

Showing 29 of 82.

Showing 31 of 82.

Showing 34 of 82.

Showing 36 of 82.

Showing 37 of 82.

Showing 40 of 82.

Showing 42 of 82.

Showing 67 of 82.

Three or more filters are set; the opportunities shown are the ones that match.

No opportunity matches these choices. Set one of these filters back to Any: Debate;Project size;Cost;Related work;How far checked;Data the design needs;Evidence from (place);