IDC Revision

Week 9 · Rodgers (2021)

Promises and Perils of AI and Advertising

Themed Issue Introduction: Promises and Perils of Artificial Intelligence and Advertising

Part of: Digital Marketing Topic 9 — AI in Advertising · Reading (editorial / themed-issue introduction) Citation: Rodgers, S. (2021). Themed issue introduction: Promises and perils of artificial intelligence and advertising. Journal of Advertising, 50(1), 1–10. Key concepts: AI Advertising, Machine Learning, Narrow AI, Algorithmic Transparency, AI Influencer

TL;DR

This editorial introduces a Journal of Advertising themed issue on AI in advertising. It proposes a working definition of AI advertising — "brand communication that uses a range of machine functions that learn to carry out tasks with intent to persuade, with input by humans, machines, or both" — and positions it as a distinct subdiscipline at the intersection of cognitive science, computer science, and advertising. Rodgers offers a classification schema (AI type × AI function × learning type), summarises the six themed-issue articles in "promises vs perils" terms, and flags algorithm bias and algorithmic transparency as the field's central emerging concern.

Why It's on the Reading List

It frames the entire AI-in-advertising topic: it defines the field, supplies the vocabulary (narrow/general/super AI, supervised/unsupervised learning, building blocks) and previews the case studies (AI influencers, in-store AI, social-media listening) the rest of the topic draws on. High-yield for definitions and the classification schema.

Background & Research Question

AI is described as the most exciting recent advance in advertising, but it is "couched within a multitude of questions": What is AI advertising, and what are its promises and perils? For consumers, perils include loss of privacy and control; for marketers, a steep learning curve and uncertain ROI. Yet as firms generate exponential data, AI becomes "less an option and more a necessity". The editorial's purpose is to stimulate interdisciplinary thinking and find common ground across the themed-issue articles.

Context statistics cited

Over 75% of consumers already use an AI-powered service/device; ~53% growth expected in AI marketing in 2021; global digital advertising forecast at $517.51bn by 2023 with AI taking ~80%.

Key Concepts & Definitions

Definition — AI advertising (Rodgers' working definition)

"Brand communication that uses a range of machine functions that learn to carry out tasks with intent to persuade with input by humans, machines, or both." Here advertising = brand communication with intent to persuade; AI = machine functions that learn with or without human help.

Definition — AI influencer (Thomas & Fowler)

"A digitally created artificial human who is associated with Internet fame and uses software and algorithms to perform tasks like humans."

Rodgers stresses AI advertising is distinct from but related to computational advertising. It can be a specific type (AI endorsers) or part of a larger process (creative AI); experienced in a specific context (social media) or a broader branded experience (in-store).

Framework / Model

A classification schema with three dimensions intended to "house" research on AI advertising (illustrative, not exhaustive — dimensions overlap, e.g. machine learning is both a function and a learning type).

1. AI type (strength of intelligence):

Type Analogy Capability
Narrow / weak AI (NAI) Infant One specific task at a time
General / strong AI (GAI) Adult Complex tasks; perception, learning, problem-solving; human senses
Super AI (SAI) Beyond human Supersedes human intelligence; not yet possible

Rodgers argues the themed-issue research is best classified as GAI (e.g. autonomous AI influencers; relatively autonomous creative-advertising systems).

2. AI functions / building blocks: natural language processing, image and speech recognition, problem solving. These define what data can be analysed and which problems addressed.

3. Learning types: machine, deep, supervised, unsupervised, reinforcement learning. (Reinforcement learning was not used by any themed-issue study.)

Main Arguments / Findings — the six articles (promises & perils)

# Article (authors) Promise Peril
1 AI influencers as endorsers (Thomas & Fowler) As effective as human celebrity endorsers for favourable brand responses A transgression can taint all AI influencers; consumers may disavow the category
2 Machine-learning ad placement / semantic relatedness (Watts & Adriano) Context-aware training improves placement efficiency Only as good as the trainers' assumptions; needs broader vocabulary
3 Creative Advertising System / CAS (Vakratsas & Wang) Road map to rethink/scale advertising creativity Hard to assess novel transformational creativity computationally
4 Detecting image–text mismatch on Instagram (Ha et al.) Computer vision opens new research on contextual congruence (452,616 posts) Substantial training needed; broader visual inputs required
5 AI-enabled in-store communication / "just-walk-out" (van Esch et al.) Stimulates patronage via sensory in-store brand communication Positive effect undermined if consumers perceive AI as a privacy threat
6 Trusting social-media-listening AI / LIWC vs human (Hayes et al.) SMLPs boost data-collection speed and research robustness Machines less nuanced than humans on brand-specific coding (10,000 Nike "Dream Crazy" eWOM comments)

Implications for Marketers

  • Treat AI advertising as a distinct subdiscipline, not a gimmick — it touches every stage of the advertising management process (research, targeting, creation, placement, performance).
  • The recurring trade-off is efficiency/personalisation vs privacy (article #5).
  • Humans may have to clean up machine mistakes — replacing a transgressing AI influencer with a human endorser can repair brand perceptions.

Exam Takeaways

Likely exam points
  • Rodgers' working definition of AI advertising (brand communication + machine functions that learn + intent to persuade + human/machine input).
  • The classification schema: AI type (narrow/general/super), AI functions, learning types.
  • Why the themed-issue work is classed as General AI.
  • "Promises and perils" framing — be able to give the AI-influencer and in-store examples.
  • The closing concern: algorithm bias and algorithmic transparency, plus the danger of assuming AI is "value-neutral and unbiased".

Summary

  • Editorial/conceptual introduction, not an empirical study.
  • Defines AI advertising and locates it at the intersection of cognitive science, computer science and advertising.
  • Provides a 3-dimension classification schema for organising AI-advertising research.
  • Frames six studies as promises/perils; ends warning about algorithmic bias and transparency.