Diving brief:
- Consumers quickly identify video ads made with generative artificial intelligence (AI) and are not enthusiastic about what they see, according to new research from NielsenIQ (NIQ).
- AI-generated creatives were consistently rated more “boring,” “boring,” and “confusing” than ads created using traditional methods. Even AI results rated as high quality did not leave as strong an impression, highlighting a gap between the content and existing memory structures.
- Low-quality AI ads still effectively communicated brand associations, but proved cognitively taxing and distracting for viewers. These results follow several high-profile advertising errors related to AI in marketing this year.
Dive overview:
Marketers appear firm in their commitment to ramping up experimentation with generative AI, but industry hype may not match consumer sentiment. The NIQ research, which was conducted using surveys as well as methodologies such as eye tracking and implicit response time, found that even the most refined AI-generated ads cause some degree of dissonance among viewers while being less memorable than conventional advertising.
The result is that AI-generated ads risk undermining the sense of authenticity – a desirable trait in marketing – while creating a negative halo effect around a brand that could dampen perceptions beyond a individual campaign. NIQ will present more on the topic at the Consumer Electronics Show in Las Vegas next month. The company did not cite any of the efforts evaluated for the report.
“Brands and agencies are innovating at a rapid pace, leveraging AI-generated content in their advertising,” said Ramon Melgarejo, president of strategic analytics and insights at NIQ, in a statement on the research. “They need to be careful, because our study reveals that consumers are very sensitive to the authenticity of advertising creatives, both at the implicit (unconscious) and explicit (conscious) level. »
The level of investment in generative AI has been astonishing, with marketing being no exception. The agencies have committed to spending hundreds of millions of dollars to improve their AI know-how as brands continue to integrate the technology into major campaigns. But AI-powered creation remains a tricky subject despite the generalization of tools like ChatGPT.
Coke’s recent holiday ads made with AI sparked controversy, as did an effort earlier this year at Toys R Us which featured the first branded video to use OpenAI’s text-to-video conversion tool, Sora. AI-generated assets began to receive negative nicknames like “slop” and the NIQ results highlight that consumers’ brains have difficulty processing what they are watching, which distracts them from a campaign’s message and causes a strange valley effect. Many AI-generated videos feature actors with limited movement and facial expressions, as well as odd editing rhythms. and a distinctive shiny appearance.
AI has some strengths, namely that it can easily surface implicit brand associations. NIQ also acknowledged that AI video tools will likely become more sophisticated, which could address some of the frustrations viewers currently experience when it comes to realism.
“Even the lowest quality AI-generated ads managed to convey the desired brand identity” the NIQ report says. “Each of these advertisements was able to reinforce at a strong level the mental network of associations of their respective brands…”
That said, in the short term, marketers may want to focus on lower-risk use cases of technology, such as early-stage ideation, storyboarding, and creative release management.
