The article argues that AI models disproportionately reward recent timestamps, sometimes over content quality, citing a Waseda University study. Tests across GPT-3.5, GPT-4o, LLaMA-3, and Qwen-2.5 showed that newer-dated pages can jump rankings dramatically and even flip relevance judgments, with model-to-model variation and a noted ChatGPT freshness flag. For HVAC, plumbing, and electrical contractors, this recency bias can impact AI Overviews and assistants, affecting visibility and booked jobs. The recommended approach is ethical: regularly refresh revenue-driving pages with real updates, add proper date metadata, and keep seasonal/local details current to capture AI-driven demand without falsifying dates.

AI Loves “Fresh” Content—Even If It’s Fake. Here’s How Contractors Win Without Lying

TL;DR AI models overvalue “newness.” A study shows fake dates can push pages up dramatically. Don’t fake it. Instead, refresh your money pages with real updates, add proper date metadata, and keep seasonal/local info current. Freshness helps your pages get pulled into AI answers—more calls, not just clicks.

What the study found (and why I care)

Waseda University researchers tested how AI models handle recency. Short version: they’re biased. Add a newer date and content looks more “relevant,” regardless of actual quality.

• Models tested: GPT-3.5, GPT-4o, LLaMA-3, Qwen-2.5. All favored newer-dated content.
• Pages with fresh timestamps jumped as much as 95 ranking positions.
• About 25% of relevance choices flipped based on the date alone.
• Top-ranked results looked 0.8–4.8 years newer on average; results ranked 61–100 looked up to 2 years older.
• Bias varied by model: LLaMA-3-8B was most biased; Qwen-2.5-72B was least; OpenAI sat in the middle.
• There’s even a config flag noted in ChatGPT—“use_freshness_scoring_profile: true”—that signals a preference for recent content.

Bottom line: AI systems tend to reward recency, sometimes over quality. That encourages a “temporal arms race” where publishers play the date game to stay visible.

Why this matters for HVAC, plumbing, and electrical

More homeowner journeys start with AI-infused search (AI Overviews, chat assistants, summaries). If your “Furnace Repair” page says “Updated 2023” and a competitor says “Updated 2025,” these systems may favor the competitor—even if your content is better. That can mean fewer placements in AI answers, less visibility, and fewer calls.

I’m not here to pump pageviews. I want booked jobs. But you only get the call if you show up. This bias means freshness signals (real or fake) can influence who gets surfaced by AI. So we play the game—ethically—and make freshness work for revenue.

Don’t fake dates. Do real refreshes that convert.

You don’t need to lie. You need a repeatable refresh cadence on the pages that drive revenue. Here’s the playbook I run for contractors:

1. Prioritize the money pages. Service pages (AC repair, furnace install, drain cleaning, panel upgrades), financing, emergency services, and top city pages. Blogs come after.
2. Quarterly refresh cycle. Touch each money page every 90–120 days. More often during peak season.
3. Make real updates:
   • Update pricing ranges and promos (“as of” month/year). Keep compliance in mind.
   • Add current codes/regulations (SEER2, refrigerant changes, electrification incentives, 25C credits).
   • Seasonal FAQs (spring AC tune-up checklist

     User-provided (no source URL)