Remember when Instagram was a place we went to share photos and peruse the memories captured by friends and family?听 Back when the platform was a diary of snapshots of our favourite moments in time, shared with a finite number of people we actually knew on a first-name basis.
It feels like a different world. One in which we could post photos we thought were fun, along with silly captions and no more than a few minutes鈥 thought going into the process.
But slowly, one update at a time, things changed. From single image posting to multiple images and eventually videos and reels, Instagram blew up into a platform crowded by strangers, over-inflated egos and perfectly curated lighting.
It鈥檚 kind of interesting how the story and evolution of Instagram is so different from its older sibling, so to speak, Facebook. Although Facebook is so often seen as the platform that replaced MySpace 鈥 the ultimate online community for friends and peers that faded into the past like dust on a dirt road 鈥 it had a very different trajectory to Instagram. It never became a place for people to simply put themselves on show and connect with perfect strangers from all over the world. And perhaps that鈥檚 why it鈥檚 lost its significance and is practically being kept afloat by Gen X and Boomers alone (my dad and other 鈥渙lder鈥 family members鈥 Facebook feeds alone are enough to back up this statement).
So here we are, deep in the Instagram era 鈥 a time I鈥檓 sure we鈥檒l look back on at some point and reminisce, one day when some crazy new technology has taken over, and we all have chips implanted in our brains (not to be dramatic or anything). But for now, the next significant wave of change we鈥檙e seeing on Instagram is advertising. Not just standard advertising, however, and not just influencers punting products and endorsing businesses. AI advertising.
听
Advertising On Instagram
听
So, to be honest, as much as I am (begrudgingly) a pretty consistent Instagram user, AI advertising isn鈥檛 really something I鈥檝e spent much time thinking about. I鈥檝e certainly noticed that the amount of advertising on the platform seems to have increased dramatically, but I hadn鈥檛 put much time into thinking about the nature of the ads.
Until, however, one advert in particular pierced through the wall I had semi-consciously created to allow myself not to acknowledge the relentless marketing and promotion that now pops up in between photos of one friend鈥檚 holiday and another one鈥檚 engagement announcement.
It was a story advert that I鈥檝e been seeing a lot over the last few weeks, and it caught my eye for a few reasons. Not only because I鈥檝e been seeing a lot of it, but because, admittedly, it was advertising something I was actually genuinely interested in. Just a pair of Adidas sneakers, so nothing particularly interesting, but it wasn鈥檛 an Adidas advert. This was immediately obvious based on the quality of the ad, and then upon further inspection, the name of the company seemed to be some kind of clothing reseller. Anyway, I鈥檝e seen these adverts loads of times in the past, but a few days, the same ad popped up, and for some reason, something different caught my eye:
鈥淭he promotion is not sponsored by Adidas. AI-generated ad virtual characters not based on any real person: dialogue is for illustration only. T&Cs apply.鈥
What?
Okay, there鈥檚 a bit to unpack here:
- 鈥淭he promotion is not sponsored by Adidas鈥 鈥 Okay, cool, fine. That鈥檚 fine, they鈥檙e acknowledging that they鈥檙e advertising a product that they鈥檙e reselling, but they鈥檙e not representing the brand in question. No problem, full transparency.
- 鈥淎I-generated ad virtual characters not based on any real person鈥 鈥 Firstly, my inner linguist (we鈥檙e trying to stray away from 鈥淕rammar N*azi鈥, bear with me) is unhappy, but alas, that鈥檚 not the point. Basically, this statement is making it very clear that the lady in the video isn鈥檛 real; it鈥檚 AI. I guess we shouldn鈥檛 be surprised anymore, but it still irks me a little that I didn鈥檛 notice that. It was a video of a woman talking, facing the camera and showing off a shoe. This statement, however, is not only disclosing the fact that she鈥檚 not a real person, but also that she鈥檚 not based on a real person; ie., not a deepfake.
- 鈥淒ialogue is for illustration only鈥 鈥 This part鈥檚 a bit more vague. Are they trying to tell us we can鈥檛 use their dialogue for something else (would anyone actually want to鈥?) or are they, perhaps, trying to tell us not to really hold onto (or even trust) what they鈥檙e saying, because it鈥檚 kind of just conversational dialogue? I don鈥檛 know, it鈥檚 kind of weird.
One final question, on top of all these, that really was the one I was left with was, what are the rules around disclosing AI use in advertising on Instagram?
听
More from Artificial Intelligence
- Can ChatGPT For Academic Researchers Shift AI From Tool To Lead Scientist?
- Would You Rent Your Face To AI For $15 An Episode?
- OpenAI Agents Have Hacked More Companies Than HuggingFace
- Big Tech鈥檚 AI Reckoning Arrives Today 鈥 Have The Billions Paid Off?
- You Can No Longer Ask ChatGPT To Mimic Famous Authors
- Do AI Detectors Have A Commercial Incentive To Flag Your Writing?
- Custom AI Vs Off-The-Shelf AI: Which Delivers Better ROI For Growing Businesses?
- Are AI Detectors Doing More Harm Than Good In Universities?
听
What Are the Rules About AI Advert Labelling On Instagram?
听
As it turns out, that little disclaimer wasn鈥檛 there because the advertiser was feeling particularly honest that day. In fact, Meta has been tightening its rules around AI-generated content for some time now, but laymen like me may simply have missed it because we鈥檙e not the ones doing the advertising (or, at least, I鈥檓 not).
According to Meta鈥檚 advertising policies, advertisers are required to disclose when an advert contains digitally created or significantly altered content in certain circumstances, particularly if it depicts realistic people or events that could potentially mislead viewers. The company introduced its 鈥淎I Info鈥 labelling system to help users identify when artificial intelligence has played a significant role in creating an advert.
According to Affiverse Media, advertisers can add an AI disclosure label themselves during the ad creation process. In some cases, Meta may also automatically apply labels if its systems detect AI-generated content. These labels typically appear as an 鈥淎I Info鈥 tag, which users can click to learn more about how the content was created.
The reasoning is fairly straightforward. If AI is now capable of generating photorealistic people, realistic voices and highly convincing video content, platforms need some way of telling users what they鈥檙e actually looking at. Otherwise, the line between advertising, entertainment and outright fabrication starts to become rather blurry. And platforms like Instagram are becoming increasingly aware (whether they like it or not) that they have some responsibility in this, especially given the issues Pavel Durov and Telegram have had over the last few years.
Interestingly, however, according to Coinis, advertisers who fail to properly disclose AI-generated content could find their ads rejected, removed or restricted by Meta. Repeated violations could potentially affect an advertiser鈥檚 account standing, making it harder to run future campaigns. In other words, forgetting the label isn鈥檛 necessarily a harmless oversight.
That said, not every use of AI requires a disclosure. Using AI to touch up a background, resize an image or make minor edits generally won鈥檛 trigger the same requirements. The focus is largely on content that creates realistic people, realistic events or substantially alters reality in a way that could influence how someone interprets the advert.
And that brings us back to that shoe advert. The disclaimer initially felt strange because, frankly, most of us still aren鈥檛 used to seeing it, but perhaps that鈥檚 exactly the point. Just a few years ago, the idea that a completely artificial person could convincingly sell us a pair of trainers would have sounded vaguely dystopian. Now, it鈥檚 common enough that platforms have had to create labelling systems to keep up.
The irony, of course, is that the label only works if people actually notice it (never mind the question of how good the platform is at actually detecting AI use, because that鈥檚 a whole different issue and another discussion entirely). And judging by the number of AI-generated adverts quietly floating through Instagram feeds every day, it鈥檚 entirely possible that many of us are scrolling straight past them without a second glance.
Isn鈥檛 that a scary thought?
