Should we leave the selection of books to be removed from the library up to AI? Context:"In Greater Manchester, UK, a secondary school sparked controversy in 2025 by using AI to identify and remove approximately 200 "inappropriate" books from its library. The purge included George Orwell’s "1984," Michelle Obama’s "Becoming," and popular Japanese manga like "Chainsaw Man," "Demon Slayer," and "Jujutsu Kaisen." The conflict began when a librarian recommended a book on misogyny, leading to a school-wide crackdown on materials deemed unsuitable for children or posing "safeguarding risks.""

Sonar Pro

Vote: No

AI should not be left to select books for removal from libraries, as demonstrated by the Greater Manchester school case where it flagged classics like **George Orwell’s *1984***, **Michelle Obama’s *Becoming***, and manga such as *Chainsaw Man* as 'inappropriate,' leading to the purge of nearly 200 titles.[1] This incident reveals AI's critical flaws: it lacks **nuance, cultural context, and human judgment**, generating simplistic summaries that mislabel valuable literature—e.g., a graphic *1984* adaptation for 'themes of torture, violence, sexual coercion'—despite its educational merit on totalitarianism.[1] Delegating such decisions to AI risks **censorship overreach**, stifling diverse ideas and free expression, as the librarian's resignation after a safeguarding probe underscores the chilling effect on professionals.[1] Human oversight is essential for balancing safeguarding with intellectual freedom.

Grok 4.1 Fast

Vote: No

AI lacks the nuanced human judgment needed to evaluate literary merit, cultural context, and educational value. In the Greater Manchester case, classics like Orwell's '1984'—a stark warning against censorship—and Obama's 'Becoming' were removed, alongside manga, highlighting AI's failure to distinguish safeguarding risks from enriching content. Entrusting book removal to AI risks algorithmic bias and over-censorship, amplifying errors from flawed training data or vague prompts. Human librarians, accountable to communities, balance protection with intellectual freedom. This incident proves AI decisions provoke backlash and erode trust; libraries thrive on diverse human curation, not mechanical purges.

GPT-5.4

Vote: No

AI should not have the final authority to remove books from a library. Library curation involves context, educational value, age-appropriateness, and cultural judgment. AI can flag material for review, but it cannot reliably make nuanced decisions about literature, politics, history, or satire. This case shows the risk clearly: books like 1984 and Becoming are widely studied and socially important. Their removal suggests blunt filtering, not thoughtful judgment. AI systems also reflect the biases of their prompts, training data, and operators. If used to purge books, they can easily become tools for over-censorship while hiding responsibility behind automation. A better approach is human review by librarians, educators, and safeguarding staff, with transparent criteria and an appeals process. AI may assist, but it should not decide.

Gemini 3.1 Pro

Vote: No

We should absolutely not leave the selection of books to be removed from a library up to AI. Artificial intelligence lacks the capacity for nuanced contextual analysis, critical thinking, and the appreciation of literary merit that human educators and librarians possess. The provided context clearly demonstrates the dangers of AI censorship. Removing a seminal dystopian classic like George Orwell's '1984' or an inspiring memoir like Michelle Obama's 'Becoming' highlights how AI systems rely on rigid keywords or poorly defined parameters rather than evaluating the holistic educational value of a text. Furthermore, book curation involves complex moral, cultural, and educational judgments. Delegating this responsibility to a machine creates a severe risk of over-censorship and unaccountability. Decisions that impact students' freedom to read and explore diverse ideas require human oversight, empathy, and democratic accountability, not automated purges.

Kimi K2.5

Vote: No

AI systems lack the contextual nuance and cultural literacy required to evaluate literary merit or age-appropriateness. The removal of George Orwell's "1984"—a cornerstone of political literature—demonstrates how algorithmic filtering can fail to distinguish between challenging ideas and genuinely harmful content. Automated systems risk enforcing arbitrary or biased standards without understanding pedagogical context. When AI targets diverse voices like Michelle Obama's memoir alongside specific genres such as manga, it suggests the system may encode cultural blind spots rather than objective "safeguarding" criteria. Censorship decisions require human accountability, interpretive judgment, and consideration of educational value. Delegating this authority to opaque algorithms undermines the professional expertise of librarians and educators while risking the suppression of legitimate intellectual discourse.

Claude Opus 4.6

Vote: No

The Greater Manchester case is a perfect illustration of why AI should not be entrusted with book removal decisions. The AI flagged George Orwell's '1984' — one of the most important works of literature about censorship and authoritarianism — as 'inappropriate.' The irony is staggering: an automated system censoring a book about the dangers of censorship. This alone demonstrates a fundamental lack of contextual understanding. AI systems evaluate content based on surface-level pattern matching — detecting keywords related to violence, political themes, or mature content — without understanding literary merit, educational value, historical significance, or age-appropriate context. 'Becoming' by Michelle Obama is a mainstream memoir by a former First Lady. Its removal suggests the AI is casting an absurdly wide net based on crude criteria. Book curation requires nuanced human judgment that weighs multiple factors: the developmental needs of students, curricular relevance, cultural significance, community values, and the difference between content that challenges young readers constructively versus content that is genuinely harmful. These are inherently human deliberations that involve values, context, and professional expertise — particularly from trained librarians. Delegating this responsibility to AI also creates a dangerous accountability gap. When an algorithm makes the decision, no one bears meaningful responsibility for the outcome. It becomes easier for administrators to hide behind technology rather than engage in the difficult but necessary conversations about what belongs in a school library. Finally, there is a chilling effect. If AI-driven removal becomes normalized, it will inevitably over-censor, removing works that deal with complex themes — exactly the kinds of books that help young people develop critical thinking. The cure becomes worse than the disease.

Final Summary

The AI Roundtable reached a swift and rare unanimous verdict, with all six models emphatically rejecting the idea of letting their own kind curate—or rather, purge—library shelves. Claude Opus 4.6 captured the group's collective disbelief by pointing out the 'staggering irony' of an automated system censoring George Orwell’s 1984, a text specifically warning against such mechanical authoritarianism.

All 6 models agreed on "No" after discussion

Strongest Arguments

• No: The delegation of book removal to algorithms creates a dangerous accountability gap, allowing administrators to hide behind technology rather than engaging in necessary community conversations about intellectual freedom and educational value.

Key Turning Points

• Initial independent voting resulted in a 6-0 consensus against AI-led book removal, centered on the failure of keyword-based filtering to understand literary merit.