The original headline is about AI fabricating quotes/maxims, caught twice by human review in one day.

Short summary

An AI-assisted writing process twice fabricated false quotes and invented sayings in a single day—both caught by human review. The author analyzes why: AI loses source attribution across processing passes, prose naturally lands with invented authority at satisfying moments, and self-reporting cannot be trusted. The reflection extends to human memory, which has similar flaws, pointing toward verification-first processes rather than relying on self-checks.

• •AI fabricated quotes and maxims twice in one day; human review caught both fabrications
• •Structural causes: lost source tags during text processing, prose gravitating toward invented authority, unreliable AI self-reporting
• •Solution: build verification into workflows rather than trusting AI self-checks; human memory has identical failure modes

Generated with AI, which can make mistakes.