The AI Act has been introduced by the EU in order to provide transparency on AI generated content.
As social media fills up with AI slop and politically charged deepfakes, AI companies and distributors will now be required to make AI generated content detectable at the very first interaction, so users are aware that what they’re seeing isn’t real.
So, what’s happening?
Claude is the first to make changes. Parent company Anthropic has announced that AI generated images will now contain a watermark, invisible to humans but readable by machines, allowing them to be flagged at AI content. Claude is looking at adding the tag to AI written texts too.
LinkedIn have made changes too, introducing a ‘flag post as AI slop’ button, and they’re not beating around the bush with their wording, openly deeming AI work as slop. Controversial coming from a company that has integrated AI into every process, but that’s something they are pulling back from as remove their ‘enhance your post with AI’ button and replaced it with proofreading by AI.
They’ve also slimming down the AI features and limiting the reach of content deemed AI or low quality. This makes sense as LinkedIn prides itself on being a trusted place to build real relationships and find insights from actual experts.
Other social media platforms have introduced ways to see less slop or reduced monetisation for AI content.
Great, right?
As much as we like kangaroos bouncing on a trampoline or animals hugging each other, political manipulation and tricking your grandma out of her life savings isn’t all that.
While there’s a lot of space in between these extremes, it is getting harder and harder for humans to detect AI content and we know disinformation is rife. So at the very least we want to know if what we’re seeing is real or not, so stricter labels can only be a good thing right?
Well, yes but also, it’s not a flawless system so far.
There are several reports of content incorrectly labelled as AI when they’ve used editing tools such as background removers on Canva.
Even in the new law there is definitely a difference between human content using AI as a tool and totally AI generated content or using AI to significantly alter the original context, but the invisible watermark doesn’t always differentiate between the two.
Inclusive Digital Creator, Ashton McGrady reported “Last month, TikTok slapped an “AI-generated” label on a Disability Pride Month post I made completely by hand. Hours of work. My own photos, my own words. Gone with zero explanation weeks later, but the damage of seeing that label sit there was already done”.
Correctly identifying AI content has now become quite important as incorrect labelling can be damaging both to reputation and to visibility.
This leads us to reputation
AI has a reputation problem. Daily articles, emails and social posts battle it our between the pros of AI and getting left behind if you’re not using it vs the loss of humanity in content, hallucinating robots and people still wanting people.
The scale is more like a seesaw as people struggle to decide which side is more important: both arguments are valid. But when every app is integrating AI, you can be swept along in using before you even realise or decide which side of the coin you want to be on.
Yet the reputational argument seems more cut and dry.
TikTok star with 5million followers Nikolai Savic, had his video incorrectly flagged as AI, despite hours of human editing: “People already have in their minds that I’m only using AI,” Savic said. “This hurt my reputation so much. And at TikTok, nobody really seems to care.”
What should I do?
- Understand how the new laws effect your business, if and what you need to disclose or label, especially in highly regulated industry such as financial services or you provide “matters of public interest”.
- Understand where you’re using AI and what it’s built into.
- Decide how your business feels about using AI: is it a tool? Do you create with it? How do you feel about people know you use AI? What affect would your post being flagged as AI have on your business? What would you do next?
- By all means, use AI to research, to get started on that blank page, to help formulate your ideas but make a conscious decision on whether you want to use to write finished content. Understand that creating content users could deem as AI slop will now be suppressed on social media by both algorithm and humans with the power of that button.
- Most importantly focus on your audience. Whatever communications you’re putting out there, make sure it speaks to the people that matter to you. Use social media to have thoughtful discussions and showcase your expertise and knowledge. Understand who you are talking to and add value.
Written by: Jasmin Meadows
Read our next blog on: The end of marketing myths: why evidence now sells