Author: Steffen Meyer, Mobile Marketing Content Specialist
Before we dive into the world of AI for marketing, let me assure you that this introduction was not written by AI. While AI has come a long way in terms of language generation and natural language processing, it still has a long way to go before it can produce writing that is up to the standards of a human writer.
The paragraph above is a lie.
I told the famous text AI ChatGPT to “write an introduction to a text about AI for marketing but explain to the reader that the introduction was not written by AI because this would have been very sloppy writing.”
Pretty good writing, I’d say. But how do you know this isn’t a lie, either?
You’ll just have to trust us not to lie to you.
Though how much can you trust AI?
The vast amount of new tools presents marketers with lots of possibilities, though the more complex they get and the more they work as a black box, the less you can control or double-check the outcome.
So when employing an AI tool, you should rate them on a trust scale and adjust your skepticism and evaluation efforts accordingly.
Low trust issues: Marketer’s little AI helpers
On this stage, you place tools that modify your own content, such as:
- Automatic subtitling
- Removing backgrounds from photos or videos
- Deleting breathing pauses in videos
- Enhancing your audio to sound recorded in a studio
- Generating profile pictures for your socials
All these little AI elves help you finish tasks that some years ago would have been tedious manual work. Of course, the result isn’t always perfect: The subtitles need some refinement or the photos still have some background left over.
However, this is fine since you can easily edit the outcome and furthermore, you don’t need to worry about where the AI got the content from since it came from you.
It’s a different issue on the next stage, which is responsible for all the AI Buzz.
Mid-sized trust issues: Generative AI
Generative AI imitates human marketers: It creates:
While this sounds magical, you should always double-check the results before showing the product to the outside world.
For example, when it comes to complex topics, you will see that AI Copywriters’ responses aren’t quite right or just blatantly wrong, even if they sound totally convincing. These outputs are called hallucinations and may put you on a wrong track which could be more work to debunk than doing the research yourself in the first place.
Furthermore, you need to check the sources the AI is citing (it could hallucinate those as well) and you should be aware that the tool could have just copied copyrighted work which may put you in legal jeopardy.
The last problem mentioned especially applies to image generators: Is it legal that the AI is trained with various products of human creators and builds on all their work without paying them anything? The jury’s still out.
Noteworthy is though, the creative tools giant Adobe claims that its Firefly image generator was “trained on Adobe Stock images, openly licensed content and public domain content, where the copyright has expired. They further say that it is designed to generate images “safe for commercial use” and that the company is working on a “Do Not Train” Content Credentials tag.
So all in all, you should face these AI tools with skepticism – sometimes it could be better to just use them for inspiration than for actually creating content.
High trust issues: Marketer’s old friends
Creating assets with AI is one aspect, since you can still control the end product, but basing the distribution of ads on AI requires even more trust.
Weirdly enough, many digital marketers already do so.
E-mail marketing or social media management tools promise you to send or publish your content at the best time, or even claim that they analyze the behavior of every single recipient so that your prospects see your output at their optimal individual time.
While you can still A/B test these recommendations by establishing control groups, it’s still a lot of trust in the system.
Even more so, many marketers let AI auctioneers of ad networks like Facebook, Instagram, Google or TikTok spend their marketing budget.
Millions of all kinds of currency are managed by these AIs every day. At the same time, the so-called privacy initiatives of these companies make it even harder to verify the data you see in ad managers.
With Meta’s Advantage+ or Google’s PerformanceMax campaigns, the ad networks leave you with less manual control but promise better results. The machine knows best.
Going even further, some marketing tools promise you to fully automate and optimize your cross-platform marketing, taking even more control away from the human and giving it to algorithms.
That’s a hell of a lot of trust marketers put in these tools already, and proving that their promises are true, requires knowledge, experience and resources – which is why spending money on consultants could be a better invest than just raising your budget in the bidding wars.
As always, be skeptical
So even with all the buzz around generative AI, the AI-assisted marketing age already began years ago and with all the new tools, it’s the same advice as with the old ones: Test them, be skeptical and try to evaluate the results through processes you can trust, so that your clients trust in you.
PS: When I told ChatGPT to “write a marketing strategy based on the Marketing Master Map for Customlytics”, it answered that “we will need to consider the four quadrants of the map: Product/Service, Customer Needs, Communication, and Distribution” – which is totally wrong.
So better check out our free Marketing Master Map yourself. It will help you more than playing around with AI tools.
💡 Knowledge sharing is at the core of what we do. Sign up for our newsletter and become part of our community on LinkedIn to learn how to make apps succeed in the competitive mobile landscape.