Friedman notes that most works now under scrutiny were written, edited and submitted before the publishing sector fully grasped how many authors used AI without disclosure.
She explains that detection software such as Pangram, which she first reported on in 2024, only began to be used widely by agents and publishers earlier this year, creating a gap between older manuscripts and newer verification methods.
According to Friedman, once publishers and agents incorporate AI‑scoring tools into their acquisition workflow, surprise accusations are likely to disappear because projects will be vetted for provenance before publication.
She adds that while lawsuits may arise over copyright eligibility for AI‑assisted works, the primary battle will centre on defining human authorship rather than on detection software itself.
Friedman concludes that the industry’s focus will shift from public scandals to internal safeguards, and that any future technical solutions—such as blockchain‑based provenance—are unlikely to be adopted due to cost and complexity.