Photo: Getty Images/fizkes
Are AI-generated ads worth the risks?
A university study determined that only half of artificial intelligence-generated ads are open about the fact they were created by bots, potentially tricking consumers into false beliefs or causing confusion and dissatisfaction.
Researchers from the University of Kansas and Florida State University analyzed 1,375 AI-generated ads found on social media, news sites, search engines and video platforms in exploring the high frequency of usage in programmatic advertising.
The primary problem with the lack of transparency is that humans must follow guidelines set forth by agencies such as the FCC and FTC when creating advertising content, but AI is not bound by such restrictions so far.
“AI is not just a passive technology anymore,” said Vaibhav Diwanji, lead author of the study and a professor of journalism & mass communications at the University of Kansas, in a press release. “It’s actively being engaged in what we think — and in a way, how we make our decisions. The process has become more automated and is taking over the role of creative content online.”
The arrival of ChatGPT, DALL-E and other advanced AI tools has sparked speculation on how the technology may impact search advertising as well as certain jobs in sectors such as journalism, higher education and software design.
Some have called out the technologies’ shortcomings in replacing humans. CNET issued corrections to mistakes made in over half of the AI-written articles recently attributed to its CNET Money team, including plagiarism.
AI-generated ad content should significantly reduce production costs, leading to speculation on whether marketing costs overall could be coming down.
According to an Adweek article, brands would likely employ AI-generated ads if engagement, ratings and reach are compelling. On the downside, marketers will likely be busy cleaning up errors and need to filter through a flood of inferior ad content hitting the marketplace.
“The internet has no shortage of mediocre content at scale,” Jon Roberts, the chief innovation officer at Dotdash Meredith, told Adweek. “This is a new version of an age-old temptation.”
- Study finds only about half of AI-generated ads only labeled as such – Eurekalert
- AI-Generated Content Is Here. Should Advertisers Buy It? – Adweek
- ChatGPT in advertising: Will clients pay the same for AI-generated campaigns? – Campaign Asia
- Is AI A Risk To Creativity? The Answer Is Not So Simple – Entrepreneur
- Tech-driven marketing strategies pick up as AR/VR and AI become more accessible – Digiday
- CNET found errors in more than half of its AI-written stories – The Verge
Discussion Questions
DISCUSSION QUESTIONS: Will AI-generated advertising content drive more benefits or chaos for marketers? What advice would you have for retailers or brands exploring AI-generated ad content?
Poll
Loading …BrainTrust
John Lietsch
Chief Operating Officer, Bloo Kanoo
Gwen Morrison
Partner, Candezent & Retail Cities Consultant
Without a human review before release, AI generated content is a risk to brands and retailers. At the very least AI ads could be out of tune with brand voice — at worst an AI ad could ruin customer trust and loyalty. AI content needs guardrails, and only increases the importance of human involvement in the creation process.
AI generated content does not exempt us from our responsibilities or guidelines. We can’t simply blame the “bot.” Retailers should continue to play with AI generated “everything” but apply the same editorial review processes they apply to their human generated content. At this stage, AI generated content is only speeding up ideation. Accept and leverage those gains and wait for the rest of the process to be improved by technology; it won’t be tomorrow but it won’t be long.
AI-generated content is already prevalent and this will only continue to be. The fact is, in many cases AI-generated content is good enough, and let’s not forget that there’s a lot of human generated ad content that’s pretty bad. I can absolutely see retailers using AI generated ad content as a first version that can then be edited by humans. The ability to create countless versions of an ad with AI give it a production edge over humans and retailers should consider using it.
What a great question! I would not take a chance on AI-generated ads. The day one goes awry it will cause a huge embarrassment, like the Southwest Airlines glitch. And this comes from an advocate of using technology to improve many (but not all) functions.
Anyone familiar with “The Uncanny Valley” will understand the risk of trusting AI-generated content without a human editor reviewing/revising before release. The “something is eerily not right” phenomenon with artificial humans is uncannily perceptive at an unconscious level. AI can empower us to make better informed decisions faster, but it is still not capable of human cognition and perception.
This is an easy one: avoid embracing this. No authenticity here, and customers will see through it.
Within the last week, I read an explanation of the problems that surfaced with ChatGPT and DALL-E. The fellow explained that the AI interface was designed to reflect humans’ thoughts and words. It was noted that is exactly what they got, unfiltered humans.
AI-generated content can be a compelling first step in the content production process, but it’s certainly not the only step. Human oversight is mandatory (and also an opportunity for us to provide differentiated value from the machines!).
“Danger, Will Robinson!” (Sorry for the ’60s reference.) AI-generated content is in chapter one of a hundred-chapter story. This is no time to go fully-operational without serious oversight.
AI-generated content can be a useful tool but never, never, use it without reviewing content for publication. We’re starting to see some horror stories bubble to the surface of individuals relying solely on these tools. It’s a huge mistake to do this.
There are so many smart use cases for AI. Creative content is not one of them. Right now, the focus should be on using AI for low-creativity tasks that are risk-free from a moral standpoint if there’s a mistake.
“…But AI is not bound by such restrictions so far.” Huh? Companies are what are bound by restrictions, so unless we’ve reached the stage where AI is creating corporations — and presumably paying all the costs related to them — I don’t see how this statement makes any sense.
My advice is simple, we seem to be in Dotcom 2.0 — 3.0? 4.0? — where a buzzword is mindlessly attached to … everything. When you see it used, run.
The best advertising reflects relationships consumers have with a categories, products and brands. AI generated content can’t deliver emotive insights that connects human situations and brand imagery. AI can select human developed content to match up with interests, preferences and purchases. That’s simply delivering personalization.
Disclosure – I work for Microsoft. When Microsoft first highlighted OpenAI’s offerings late in 2022, it was positioned as a “co-pilot” to aid coders writing code. At that point it seemed logical that its use would likely expend to other areas and with it the point about being a “co-pilot” or assistant remained relevant.
A few years ago Gartner talked about AI not standing for Artificial Intelligence but “Augmented Intelligence” and this is how these developments must be thought of — something to assist the human — to augment and to help — to make faster but not replace. With this in mind, the need to put careful guardrails around how AI driven content is used is critical — it must be reviewed by a human and the human must be ultimately responsible for the execution, publication and use.