A voting platform where you can have your say on the quality of today's advertising
CLUB   |     |   Sign up

Artificial intelligence in advertising: Do ads owe us truth?

No humans were harmed in the making of this ad.

(But was humanity?)

Seeing is no longer believing
There's a lot of talk around fake news. It seems we're now heading towards fake advertising as some 'creative' uses of artificial intelligence (AI) show us increasingly sophisticated - i.e.hard-to-spot - illusions.
Backstory examples: a 2018 anti-malaria campaign showing footballer and style icon David Beckham apparently speaking nine languages, and supermodel Bella Hadid apparently kissing another woman in 2019 (above).
Not to mention the 2018 Lexus car ad entirely scripted by AI (below).
Free of rights
Ya, but what's new?
Well, one recent development of interest is 'worry-free' advertising through the use of AI-generated people.
Eh?
Fresh young start-ups such as Generated Photos are mashing up human features to create startlingly lifelike images that have the advantage of being free of rights.
Hm.
Are you thinking the same? Depends on what kind of rights, right?
The female reporter (pictured) for this 2020 CBC News report said she was unable to distinguish 100% between real people and the AI-generated images such as the one shown here.

"It's a little bit of digital Frankensteining"

Tech columnist Ramona Pringle
Such images, explains Pringle, could be used in any situation where you might normally use stock photos.
The advantages are obvious, particularly for small/no-budget advertisers, or for those with less attractive products which might struggle to find real models, let alone celebrity endorsements.
As Generated Photos points out:

"All images can be used for any purpose without worrying about copyrights, distribution rights,
infringement claims, or royalties."

Do ads owe us truthfulness?
The sliver of Joe Public we canvassed on this didn't in general object to AI in advertising, but did warn against 'unintended consequences' of a good thing going too far, such as reduced trust - surely the last things advertisers want more of now or in the future?
Respondents were also largely unaware of the illusions created by AI as in the above examples, and suggested the use of transparency disclaimers to avoid fooling their audiences.
"One concern is whether ads owe us truthfulness," points out Carissa Véliz, Associate Professor at the University of Oxford, and author of the book
´Privacy Is Power'.
"If they do (and I think we all agree they do, to one extent or another), then do AI-generated images count as truthful?
What if we label them as such? Is that enough?
Suppose we see beautiful people created by AI, and we know they are created by AI. Can we still have bad consequences from that (e.g. people not accepting
their own imperfect bodies), even if we know these people are not real?
At the moment, we know ads are photoshopped to look perfect (perfect skin, perfect figure), and that still has bad effects on people."

(Carissa Véliz, Institute for Ethics in AI)
Perhaps the most successful outcome for all concerned is just to stick with entirely natural outcomes, as in footballer/broadcaster Gary Lineker evolving into KFC's Colonel Sanders....



Posted by Tree Elven on 07/08/2021

Keywords: Artificial intelligence in advertising, Carissa Veliz, AI ads, fake advertising, AI generated images

Other Blog Posts

Get our newsletter and nothing else! No spam or third party mailings.