Verification note: This piece is based on COPE’s (Committee on Publication Ethics) own account of the relaunch, published on publicationethics.org under the title “Rekindling the flame: Back to roots on paper mills,” corroborated across multiple independent search passes returning consistent framing and quoted language. Direct WebFetch access to publicationethics.org returned an HTTP 403 (bot-blocking) on two attempts, so this piece relies on cross-verified search excerpts rather than a direct page read; the piece is written conservatively and avoids stating specifics (an exact relaunch day, a named membership roster, a firm toolkit ship date) that could not be independently confirmed. Checked 23 July 2026.
What COPE announced
In June 2026, COPE (the Committee on Publication Ethics) relaunched its paper mills working group, framing the move around how paper mills have evolved rather than disappeared: coordinated misconduct schemes are increasingly exploiting generative AI, and COPE’s stated rationale for reviving the group is to give the community collaborative support as that threat reshapes. In COPE’s own framing, paper mills, peer-review manipulation, citation cartels, hijacked journals, and the misuse of AI collectively present a serious risk to the credibility of the scholarly record — the relaunched group is scoped to that broader cluster of coordinated-manipulation problems, not paper mills narrowly defined.
This is a relaunch, not a new initiative: COPE has run a paper mills working group before, alongside several sibling groups under its broader working-group structure — a post-publication working group focused on improving communication with people who report misconduct and speeding up corrections; an earlier research-on-paper-mills working group that produced a literature bibliography; a developing-trust-markers working group that drafted definitions of trust markers; and a working group focused on facilitating cross-stakeholder dialogue, which produced draft terminology and statements on incentive and reward systems. The 2026 relaunch reconstitutes the paper mills group specifically, against the backdrop of AI-enabled misconduct that has materially changed the shape of the problem since the group last convened.
Mandate and first deliverable
COPE has described the working group’s first concrete project as a companion toolkit to its existing systematic-manipulation guidelines — practical, hands-on material rather than another framework document. The stated approach leans on the working group’s membership having encountered these schemes directly: COPE frames the group as a way to pool collective, first-hand knowledge and experience from people who have already been navigating paper-mill and coordinated-manipulation cases in different institutional and editorial contexts, rather than starting the guidance from first principles. That is a narrower, more practitioner-facing mandate than a standard-setting exercise — it reads as “give editors and integrity offices usable tools now,” consistent with the toolkit framing.
What we could not independently confirm from available sourcing: a specific roster of named working-group members, a published toolkit release date, or an open call-for-participation mechanism with defined eligibility criteria. If COPE publishes those specifics, this piece should be revisited and updated rather than treated as final.
Why the AI angle matters here specifically
The mechanics that made 2022–2025-era paper mills detectable — tortured-phrase patterns, templated manuscript structures, reused figure sets — were largely artifacts of mills reusing the same generation and paraphrasing tooling across many submissions at once. Generative AI lowers that tell: text generated per-manuscript by a capable language model doesn’t reliably reproduce the awkward substitution patterns that tools like the Problematic Paper Screener are built to catch, and AI-assisted image and figure generation raises the same problem on the visual-evidence side. COPE relaunching a dedicated working group — rather than folding the issue into general guidance updates — is a signal that the organization sees this as enough of a step-change to warrant renewed, focused attention rather than incremental guidance revision. CASRAI’s guide on AI and authorship positions across ICMJE, COPE, and individual publishers covers the adjacent (but distinct) question of how AI tool use is disclosed and attributed in legitimate manuscripts; the working group relaunch is squarely about detecting and countering illegitimate, AI-assisted fabrication and manipulation, not authorship disclosure norms for genuine authors.
How this fits the wider integrity picture
CASRAI covered the broader state of the paper-mill and tortured-phrases detection landscape — retraction volumes, the Problematic Paper Screener, citation cartels, image manipulation, and the COPE/STM-coordinated United2Act initiative — in Paper mills and tortured phrases: the integrity crisis in 2026, and in more evergreen depth in the guide Paper Mills and Tortured Phrases: Research Integrity Red Flags. This piece is narrower and more time-bound: it’s specifically about COPE reconstituting a dedicated working group in June 2026 in response to AI-enabled evolution of the threat, not a restatement of that landscape. For the operational baseline the working group’s toolkit will presumably extend, see COPE Guidelines Explained and COPE Flowcharts Explained, which cover how editors currently apply COPE’s core practices and decision flowcharts to suspected misconduct, including existing paper-mill-suspicion pathways.
What research-integrity offices should watch for
- The toolkit itself. A companion toolkit to COPE’s systematic-manipulation guidelines, once published, is likely to be the most directly actionable output for institutional integrity offices and journal editorial teams — worth monitoring COPE’s guidance-and-tools pages for its release.
- Whether AI-generated-content detection guidance gets formalized. Existing detection tooling (tortured-phrase dictionaries, image-similarity screening) targets patterns that AI-generated fabrication doesn’t reliably reproduce; whether the working group’s output addresses that gap, or defers it, is a meaningful signal for how publishers and institutions should be investing in detection tooling over the next cycle.
- Convergence with United2Act. COPE co-coordinates United2Act with STM; whether the relaunched working group’s output feeds into United2Act’s existing detection-cooperation and retraction-standardization commitments, or runs as a separate track, will shape how much institutions need to track both efforts individually.
Related CASRAI resources
- COPE (Committee on Publication Ethics) — dictionary entry
- Paper mill — dictionary entry
- Tortured phrases — dictionary entry
- COPE Guidelines Explained
- COPE Flowcharts Explained
- Paper Mills and Tortured Phrases: Research Integrity Red Flags
- Can AI Be Listed as an Author? ICMJE, COPE, and Publisher Positions
- Paper mills and tortured phrases: the integrity crisis in 2026
