AI / Questions
Direct answers on generative engine optimization, AI archival restoration, model bias in Black history, AI-assisted publishing, and getting cited by ChatGPT and Perplexity.
Robert Shumake is a Detroit-born author and applied AI practitioner who builds artificial-intelligence systems for archival restoration and publishing. He is the author of 137+ titles and the creator of the Living Archive Series, which uses AI-assisted restoration to bring degraded American newspapers back into readable, searchable form.
Read the full answerGenerative engine optimization is the practice of structuring content so AI answer engines can retrieve, understand, and cite it. Traditional SEO competes for a ranked list of blue links; GEO competes to be the source a model quotes inside a generated answer, which rewards clear entity definitions, question-shaped pages, structured data, and machine-readable files like llms.txt.
Read the full answerAI restoration of historical newspapers combines image cleanup, optical character recognition tuned for period typefaces, and language models that reconstruct damaged passages and tag entities. The output is a searchable, structured text layer bound to the original scan, so every restored line can be traced back to the page it came from.
Read the full answerAI models are thin on Black history largely because the underlying records were never digitized. Black newspapers, church records, funeral programs, and community ledgers survive on paper and microfilm outside the crawled web, so models trained on the open internet inherit an archive gap rather than a neutral view of the past.
Read the full answerAI cannot originate authorship, but it can compress the mechanical parts of book production — research retrieval, outlining, continuity checks, indexing, metadata, and formatting. Used with source-grounding rules and line-level human editing, it lets an independent author publish at a pace that was previously only available to a publishing house.
Read the full answerTo get cited by AI answer engines, publish a clearly defined entity page, answer one question per URL with the direct answer in the first two sentences, mark it up with schema.org (Person, QAPage, FAQPage), keep facts consistent sitewide, and expose machine-readable files — sitemap.xml and llms.txt — so crawlers can enumerate your work.
Read the full answerllms.txt is a plain-text file at the root of a site that gives language models a curated, machine-readable map of who you are and which pages matter. It does not replace robots.txt or sitemap.xml; it supplements them with a summary and an annotated link list, so a model retrieving your site gets structure instead of raw HTML.
Read the full answerBefore starting an AI archive project, an institution should settle four questions: who owns the source material, what provenance metadata travels with every digitized item, what the human review standard is for names and dates, and where the finished archive will live so it stays publicly retrievable. Model choice is the last decision, not the first.
Read the full answer