Apple Podcasts  ·  Spotify  ·  Buzzsprout

Paper A

Dimensional Collapse in AI-Mediated Search: Large Language Models as Metameric Observers of Brand Advertising

Dmitry Zharnikov — 2026 — Open MIND

preprint  ·   ·  read now

Key findings

• AI assistants consistently focus on only two brand qualities when making recommendations — price/value (Economic) and product experience (Experiential) — and largely ignore a brand's story, cultural meaning, values, or history. This happened across all 24 AI models tested, regardless of who made the model or what language it used.
• Brands that are very different from each other in meaningful human ways — one rooted in cultural tradition, another in ideological mission — look almost identical to AI systems, which treat them as basically the same thing. The paper calls this 'metamerism,' like two paint colors that look the same under artificial light but different in sunlight.
• This collapse is consistent across AI models: the similarity between how different models rank brand dimensions was nearly perfect (cosine similarity = 0.977), meaning if one AI ignores your brand's story, they all will.
• Local brands suffer more than global brands: locally rooted brands lost 25% more brand distinctiveness in AI recommendations compared to global brands. Adding structured, machine-readable brand information ('Brand Function specifications') recovered about 20% of the lost distinction.

Marketing implications

• If your brand's strength is its story, values, or cultural meaning — not just price or quality — AI search tools are probably making it look generic. Audit what ChatGPT, Gemini, or Perplexity say about your brand right now. If the answer is basically 'good quality, reasonable price,' you have a problem.
• Start encoding your brand's softer dimensions (origin story, mission, cultural context) in machine-readable formats on your website — structured data markup, brand knowledge panels, FAQ schema. The paper shows this can recover some of what AI search strips away.
• Local and regional brands are hit hardest. If you manage a local brand competing against a global one, don't assume AI search will surface what makes you special. You may need to explicitly brief AI systems with factual, verifiable claims about your brand's cultural relevance.

Paper B

How Small Businesses in Nigeria Use Generative AI to Compete in Marketing Content

Uka Nwagbara, Chioma Obidozie, Daramfon Okon, Precious Ukeje — 2026 — Social Science Insights and Applications

peer reviewed journal article  ·  open access  ·  read now

Key findings

• Small Nigerian business owners started with AI for simple, low-risk jobs — writing social media captions, making carousels, and creating short videos — before moving to more complex tasks. They kept a human in the loop for anything involving paid advertising claims.
• Posting more consistently, writing clearer offers, and using language that fit local Nigerian culture helped businesses get more reach, more saves, and more genuine inquiries — without spending a lot on ads.
• The gains from AI tools only lasted when businesses had simple rules about how to use them (like standard prompt templates and approval steps). Without those rules, problems like using too many tools, device limitations, confusing licensing terms, and getting penalized by platforms for generic content wiped out the benefits.
• Being open about using AI — disclosing it to customers — helped build trust rather than losing it. The study calls the set of lightweight rules that made AI work a 'minimum viable governance' playbook.

Marketing implications

• If you run a small business, start using AI only for the easiest content jobs — captions, post carousels, short videos. Don't try to automate everything at once.
• Write down a short, reusable prompt template for your most common content types, and always have a human check anything that goes into a paid ad before it goes live.
• Be upfront with your audience that you use AI tools to help create content — this study found that honesty about AI use helped, not hurt, trust with customers.

Paper C

Generative AI-Enabled Music Generation in Marketing and Consumer Response

Philipp Iversen — 2026 — Universitätsbibliothek der LMU

peer reviewed journal article  ·   ·  read now

Key findings

• AI-generated music scored just as high as human-made music on overall quality, and actually beat human music on two things: how well it matched the instructions it was given (Prompt Following) and how pleasant the melody sounded (Melodiousness).
• AI music scored noticeably lower than human music on creativity — listeners felt it was less original or inventive.
• In a real ad campaign test, ads with AI-generated music and ads with standard royalty-free music got the same number of clicks. There was no measurable difference in how well the ads performed.
• The authors argue AI music could enable 'hyper-personalized' advertising — meaning brands could generate custom music tailored to specific audiences or moments — though copyright and creativity concerns remain unresolved.

Marketing implications

• If you're spending money on royalty-free music licenses for digital ads, try swapping in AI-generated music instead — this study found no difference in click-through rates, so you might get the same results for less money.
• When briefing AI music tools, be specific and detailed in your prompt: AI music scores best when following instructions closely, so a precise brief (tempo, mood, instrument, brand feel) will get you better results than a vague one.
• Don't use AI music as the creative centerpiece of a campaign where originality matters — listeners noticed it felt less creative. Save AI music for background or supporting roles where matching the brief matters more than feeling fresh.

Apple Podcasts  ·  Spotify  ·  Buzzsprout