First-pass research briefing, not a final academic review. Always read the original paper before citing.
Paper A
Party Equalization or Normalization Through Visual Generative AI in the 2025 German Federal Election
Simon Kruschinski, Fabio Votta — 2026 — Media and Communication
peer reviewed journal article · · deep dive
https://doi.org/10.17645/mac.11859
Key findings
- Smaller, less-funded parties used AI-generated visuals at higher rates than large established parties — suggesting AI tools are genuinely lowering the cost barrier to producing professional-looking campaign images.
- AI-generated posts got more user reactions (likes, shares, comments) than non-AI posts across the board. But this engagement boost was roughly the same for big and small parties — it didn't give small parties a special competitive edge.
- Most mainstream major parties labeled their AI-generated content as such, while minor parties and the far-right AfD generally did not disclose AI origins — creating a clear transparency gap.
- The AfD stood out as the only major party using highly realistic AI-generated images of people, including depictions of crime and negative emotional tone — a style consistent with populist messaging tactics.
Marketing implications
- If you work in political campaigns or advocacy, AI image tools are now cheap enough that even under-resourced teams can produce polished visuals — but don't expect that alone to close the gap with better-funded rivals who also get the same engagement boost.
- Labeling AI-generated content builds trust: established parties that disclosed AI origins did so more consistently. If you're running branded content, proactively tagging AI visuals may become a credibility signal — start doing it now before it becomes a regulatory requirement.
- Realistic AI imagery paired with negative or emotionally charged messaging drives engagement — but this is a double-edged tactic. Brands using photorealistic AI content should be aware that audiences and regulators are starting to scrutinize this approach, especially in high-stakes contexts.
Paper B
Agent-Oriented Transformation of Marketing Functions in the Generative AI Era: Introducing the Marketing Agent Loop
Merve Kadriye Yurdabak — 2026 — Dokuz Eylül Üniversitesi İşletme Fakültesi Dergisi
peer reviewed journal article · open access · watchlist
https://doi.org/10.24889/ifede.1788481
Key findings
- The paper proposes the Marketing Agent Loop (MAL): a cycle in which AI agents continuously sense market signals, generate content or decisions, interact with customers and systems, and learn from results — then start again. Think of it like a thermostat that not only adjusts the temperature but also rewrites the rules about what temperature is ideal.
- AI agents are not just tools that speed up existing tasks — the paper argues they change *who* (or what) makes decisions inside a marketing team. Pricing, promotions, customer relationships, and research can all have AI agents as active participants, not just assistants.
- The framework says this shift is structural, not cosmetic: it changes how customer value is created, how marketing decisions get made, and how organizations build new skills over time.
- The paper stresses that human oversight and ethical rules (like fairness and transparency) are essential — AI agents working without guardrails could cause real harm, so someone still needs to be in charge.
Marketing implications
- Use this framework as a mental map when deciding where to plug AI agents into your marketing workflow. The four steps — sense, generate, interact, learn — give you a checklist: is your AI actually learning from campaign results and feeding that back into future decisions, or does it stop after generating content?
- When pitching AI adoption internally, use the paper's argument that AI agents are structural participants, not just productivity tools. This helps justify bigger investments in integration (connecting AI to your CRM, pricing system, or research tools) rather than one-off point solutions.
- Build in human review checkpoints at each stage of the loop. The paper is explicit that human oversight is non-negotiable — so if you deploy an AI agent for pricing or promotions, assign a person responsible for auditing its decisions regularly.
AI & Marketing Research Radar — Big Plans Media — 2026-07-06