{"paper":{"title":"Ceci n'est pas une explication: Evaluating Explanation Failures as Explainability Pitfalls in Language Learning Systems","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"AI explanations in language learning tools often look helpful but contain flaws that can reinforce errors and erode trust.","cross_cats":["cs.AI"],"primary_cat":"cs.HC","authors_text":"Ben Knight, Danielle Carvalho, Isaac Pattis, James Edgell, Wm. Matthew Kennedy","submitted_at":"2026-04-28T22:05:57Z","abstract_excerpt":"AI-powered language learning tools increasingly provide instant, personalised feedback to millions of learners worldwide. However, this feedback can fail in ways that are difficult for learners--and even teachers--to detect, potentially reinforcing misconceptions and eroding learning outcomes over extended use. We present a portion of L2-Bench, a benchmark for evaluating AI systems in language education that includes (but is not limited to) six critical dimensions of effective feedback: diagnostic accuracy, awareness of appropriacy, causes of error, prioritisation, guidance for improvement, an"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"AI-generated explanations that appear helpful on the surface but are fundamentally flawed, increasing the risk of attainment, human-AI interaction, and socioaffective harms.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the six listed dimensions capture the critical failure modes and that the resulting pitfalls actually produce the claimed harms in real learner interactions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"AI explanations in language learning often fail across six dimensions like diagnostic accuracy and self-regulation support, creating hidden risks that demand better evaluation frameworks such as L2-Bench.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"AI explanations in language learning tools often look helpful but contain flaws that can reinforce errors and erode trust.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"df0a42c108a59b7b047251f0a6874af7356749724b1f00fcc9b7dd34af37dab4"},"source":{"id":"2604.26145","kind":"arxiv","version":2},"verdict":{"id":"38193017-0bed-43c4-bf79-af4c2e3a4a00","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-07T15:05:15.251367Z","strongest_claim":"AI-generated explanations that appear helpful on the surface but are fundamentally flawed, increasing the risk of attainment, human-AI interaction, and socioaffective harms.","one_line_summary":"AI explanations in language learning often fail across six dimensions like diagnostic accuracy and self-regulation support, creating hidden risks that demand better evaluation frameworks such as L2-Bench.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the six listed dimensions capture the critical failure modes and that the resulting pitfalls actually produce the claimed harms in real learner interactions.","pith_extraction_headline":"AI explanations in language learning tools often look helpful but contain flaws that can reinforce errors and erode trust."},"integrity":{"clean":false,"summary":{"advisory":0,"critical":1,"by_detector":{"doi_compliance":{"total":1,"advisory":0,"critical":1,"informational":0}},"informational":0},"endpoint":"/pith/2604.26145/integrity.json","findings":[{"note":"Identifier '10.1007/s00146-023-01753-4' is syntactically valid but the DOI registry (doi.org) returned 404, and Crossref / OpenAlex / internal corpus also have no record. The cited work could not be located through any authoritative source.","detector":"doi_compliance","severity":"critical","ref_index":31,"audited_at":"2026-05-19T20:26:45.548676Z","detected_doi":"10.1007/s00146-023-01753-4","finding_type":"unresolvable_identifier","verdict_class":"cross_source","detected_arxiv_id":null}],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T03:34:18.519966Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T20:26:45.548676Z","status":"completed","version":"1.0.0","findings_count":1}],"snapshot_sha256":"8580a2a05a20619e57efcf17217b415a22daa23f944bcca0e61cc20d777e3b65"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}