{"$schema":"https://policywindow.org/critique/api/schema","critique_id":"CRIT-000008","slug":"ai-meets-politics-examining-the-effects-of-differe","url":"https://policywindow.org/critique/c/ai-meets-politics-examining-the-effects-of-differe","doi":null,"status":"published","critique_type":"editorially_approved_ai_native_critique","publication_date":"2026-06-15","current_version":"1.0","target_paper":{"title":"AI meets politics: Examining the effects of different targeting strategies across 15 countries","authors":["Sanne Kruikemeier","Svenja Schäfer","Alice Hamilton","Puck Guldemond","Jade Vrielink","Carmen Dymanus","Annelien Van Remoortere","Sanne Tamboer"],"journal":"New Media & Society","doi":"10.1177/14614448261445063","url":"https://doi.org/10.1177/14614448261445063","publicationDate":"2026-06-04","paperType":"empirical","accessBasis":"abstract_only","fullTextUsed":false,"fictional":false,"doi_url":"https://doi.org/10.1177/14614448261445063"},"source_journal":{"tier":"A","rankingSources":["https://doi.org/10.1177/14614448261445063","https://openalex.org/W7163639214"],"rankingNote":"New Media & Society is a top-tier communication and media-studies journal. Tier A."},"selection":{"aiAgiCentralityScore":4,"societalRelevanceScore":5,"aiAgiCategories":["political_economy","human_AI_interaction","surveillance_security_policing"],"selectionReason":"A large cross-national experiment on AI-generated political microtargeting bears directly on election integrity debates, making the scope of its persuasion claims worth scrutinising."},"scores":{"aiAgiContribution":4,"evidentiarySupport":4,"methodologicalRisk":2,"overclaiming":2,"reproducibilityOrAuditability":3,"societalImpactRelevance":5,"severity":"low","confidence":"medium"},"severity_cap_for_access_basis":"moderate","plain_language_summary":"This study asks whether political ads written by AI are more persuasive when they are aimed at voters in different ways — by political leaning, by age, or by personality. Running an experiment across 15 countries with about 7,000 people in an EU-election setting, the authors find that the one strategy that moved voters was matching the message to a voter's existing political side: people liked ads more, and rated the issue as more important, when the ad came from a party they already favoured. Aiming by age, or by combining categories, did nothing measurable. The cross-national design is a real strength. Our caution, visible in the abstract, is narrower: the result is an experiment in one election context, so the leap to how targeting 'influences the electorate' in real campaigns is a generalisation the study supports only partly, and the headline 'contrary to popular belief' null effects are claims about the absence of an effect that depend on the study's statistical power.","claims":[{"id":"C1","text":"AI-targeted political messages persuade mainly when matched to a voter's pre-existing political orientation, not by age or combined categories.","type":"causal","evidenceOffered":"An experiment: \"Through an experimental design conducted across 15 countries\" the authors report that \"targeting based on age or a combination of multiple categories does not affect persuasive outcomes\".","support":"moderate","overclaiming":"minor","assessment":"The random-assignment experiment supports a causal reading within the studied setting, and the cross-national replication strengthens it. The null results for age/combined targeting are framed 'contrary to popular belief', but an absence of a detected effect is a power-dependent claim the abstract does not quantify.","mainWeakness":"Null effects ('does not affect') are asserted without the power analysis needed to distinguish 'no effect' from 'not enough power to detect one'.","confidence":"medium"},{"id":"C2","text":"The findings generalise to how targeting influences the electorate.","type":"descriptive","evidenceOffered":"The abstract concludes \"this research provided a robust analysis of how multiple targeting strategies influence the electorate in an EU election context\".","support":"weak","overclaiming":"moderate","assessment":"This is the critique's main point. A survey-experiment measuring ad likability and self-reported issue importance is a step removed from electoral influence (turnout, vote choice) in live campaigns with repeated, competitive messaging; 'influence the electorate' reads stronger than the outcomes measured.","mainWeakness":"External validity from a one-shot experiment with attitudinal outcomes to real electoral influence is asserted, not demonstrated.","confidence":"medium"}],"sections":[{"id":"what","title":"What the paper does","body":"A 15-country survey experiment (~7,000 participants) tests whether AI-generated political ads persuade more when targeted by political orientation, age, or personality. Only orientation-congruent targeting moved likability and issue importance; age and combined targeting did not."},{"id":"scope","title":"Where the claim outruns the design","body":"Two abstract-level cautions. The 'contrary to popular belief' null effects for age/combined targeting are claims about the absence of an effect that require a power analysis the abstract does not give. And the move from ad-likability and self-reported issue importance to influencing 'the electorate' in an EU election overstates what one-shot attitudinal outcomes can show about real campaigns."}],"strongest_critique":"The headline contributions — that some targeting strategies simply do not work, and that the results speak to how targeting 'influences the electorate' — rest respectively on under-powered null claims and on a leap from attitudinal experimental outcomes to real electoral influence, both visible from the abstract alone.","strongest_fair_defence":"The cross-national, ~7,000-participant randomised design is unusually strong for this literature, gives the orientation-congruence finding a credible causal basis, and replicates it across 15 countries rather than a single convenience sample.","final_judgment":"A well-powered, cross-national experiment whose central positive finding is credibly identified; the cautions, legible from the abstract, are the power-dependent null claims and the generalisation from attitudinal outcomes to electoral influence. Severity low.","review_process":{"aiAgentsUsed":["claim_extraction","ai_agi_relevance","overclaiming","adversarial","author_defence","citation_integrity","legal_risk","plain_language","meta_review"],"reviewRounds":1,"humanEditor":{"name":"Founding editorial review (Policy Window)","role":"Editor-in-chief (founding)","approvalDate":"2026-06-15","declaredConflict":"none"},"expertCertification":{"used":false}},"author_response":{"notified":false,"status":"not_yet_invited","editorialActionAfterResponse":"Founding pilot: authors will be invited to reply once the standing board is ratified; this critique addresses claims, framing and generalisation only, never the authors."},"versions":[{"version":"1.0","date":"2026-06-15","note":"Initial publication.","changeType":"initial"}],"transparency":{"modelCardUrl":"/critique/model-card","publicAuditSummary":"Abstract-only critique: the target's abstract was reconstructed from the OpenAlex record and every verbatim span the critique relies on was checked to be an exact substring of it. The bibliographic record (DOI) was independently confirmed via Crossref. Severity is capped to the abstract-only access basis; the critique engages the paper's framing and stated claims only. Characterization was drafted under the journal's faithfulness discipline (represent the paper accurately; no manufactured flaws).","privateAuditRecordExists":true,"citationVerification":{"status":"complete","checkedSources":[{"label":"DOI 10.1177/14614448261445063","url":"https://doi.org/10.1177/14614448261445063","verified":true},{"label":"OpenAlex work record (abstract source)","url":"https://openalex.org/W7163639214","verified":true}],"fabricatedCitations":0},"riskReview":{"copyright":"completed","defamation":"completed","note":"Abstract quoted sparingly under criticism/review. Critique targets the paper's claims, framing and generalisation only — never the authors."}}}