Apple Podcasts  ·  Spotify  ·  Buzzsprout

Paper A

The Impact of AI-Driven Marketing on Gen Z Consumer Buying Decisions: Helpful or Creepy

Sahil Khan, Ruhiya Tabassum — 2026 — International Journal of Novel Research and Development

peer reviewed journal article  ·   ·  use cautiously

Key findings

• Gen Z consumers generally respond positively to AI-powered personalised recommendations — relevant suggestions that feel tailored to their interests tend to increase their likelihood of making a purchase.
• When personalisation goes too far — for example, when ads feel like the platform has been 'listening' to private conversations — Gen Z consumers feel surveilled and their trust drops significantly.
• Privacy concerns are a strong predictor of distrust: the more worried a Gen Z consumer is about how their data is being used, the less likely they are to buy from that brand.
• Transparency matters: when AI marketing was perceived as clear and honest about how it works, Gen Z consumers reacted more positively. When it felt opaque or 'too much,' their buying intent fell.

Marketing implications

• If you target Gen Z with personalised ads, add a brief explanation of why they're seeing the ad (e.g., 'You're seeing this because you browsed sneakers last week'). Transparency reduces the 'creepy' feeling and builds trust.
• Audit how often you're retargeting the same Gen Z user with the same product. Showing ads too frequently — especially for products they barely glanced at — risks feeling invasive and can actually push them away from buying.
• If you collect data from Gen Z users, make your privacy policy human-readable and put a plain-language summary where they can actually see it. Gen Z is more likely to trust brands that are upfront about data use.

Paper B

Who Trusts AI with Their Emotions? Trust Formation and Sociodemographic Variation in LLM Use for Emotional Support

Natalia Amat-Lefort, Mert Yazan, Amanda Cercas Curry, Flor Miriam Plaza-del-Arco — 2026 — arXiv

preprint  ·   ·  test this week

Key findings

• Three things increase trust in emotional AI — feeling like the AI understands you emotionally (humanlikeness), believing your data is private, and feeling like the AI is personalized to you. One thing destroys trust: believing the AI is biased or unfair.
• Not everyone needs to trust an AI before using it. Older adults and lower-income users often skip the trust question entirely — they just care whether the tool is available 24/7 and won't judge them. Higher-educated and wealthier users, by contrast, won't use emotional AI unless they trust it first.
• Women's trust is more sensitive to privacy than men's. If a woman isn't confident her data is protected, she's much less likely to use an emotional AI tool — more so than for men.
• Users in the UK and USA responded more positively to AI that felt human-like (warm, empathetic, conversational) than users in continental Europe, who were more skeptical of that quality.

Marketing implications

• If you're marketing an AI-powered emotional or mental wellness product to women, lead with privacy — be explicit about what data you store, who can see it, and how it's protected. Vague reassurances won't work as well as concrete privacy controls.
• If your audience skews older or lower-income, don't spend your onboarding flow building trust through credentials or safety claims. Emphasize practical benefits instead: always available, no judgment, no waiting room, no cost barrier. That's what converts this segment.
• If you're launching an emotional AI product in the UK or US, you can lean into warm, human-sounding AI personas. In continental Europe, that same approach may feel manipulative — test with a more straightforward, less 'human' tone in those markets.

Paper C

PILA: Plug-and-Play Insertion for LLM-native Advertising

Zhaowei Zhang, Yuhan Fu, Yihang Zhang, Xiaohan Liu et al. — 2026 — arXiv (Cornell University)

preprint  ·   ·  test this week

Key findings

• PILA can slip ads into AI chatbot answers after the answer is already written — without touching the chatbot itself. It works like a filter that reads the finished response and rewrites it to include a relevant sponsor mention.
• Compared to other methods of inserting ads into AI responses, PILA produced better results on both sides: the answer stayed useful and natural for the user, and the ad got proper visibility. It beat prompt-only methods by about 34%, sampling-based methods by about 47%, and fine-tuning-based methods by about 8% on a combined quality score.
• Adding PILA as a plug-in improved seven different commercial AI models (like GPT and Claude) by 17–18% on the combined user + advertiser quality score, without any changes to those models.
• PILA includes a dial that lets you control how 'pushy' the ad feels — turn it up for more ad exposure, turn it down for a more natural, subtle mention — giving publishers a way to price different levels of ad prominence.

Marketing implications

• If you run a chatbot or AI assistant and want to monetize it with sponsored placements, PILA's approach shows you don't need to rebuild your AI — you just add a lightweight rewriting layer on top that inserts the ad after the main answer is done.
• If you're an ad tech builder, the 'intensity dial' concept is worth stealing: give advertisers a slider from subtle brand mention to more explicit promotion, and charge more for higher intensity placements.
• Be cautious before deploying anything like this — the paper doesn't address whether users need to be told the response contains paid content, which is a real legal risk in many markets (FTC rules, EU regulations). Get legal advice before launch.

Apple Podcasts  ·  Spotify  ·  Buzzsprout