Jasmin Bedir: Ethical AI, the great oxymoron emerging in martech

By Jasmin Bedir, CEO, Innocean

Raise your hand if you are concerned about the scarily rapid dehumanisation in all aspects of our lives, from audio deepfakes mimicking school principals and presidents, to dating AI profiles on dating apps, which could mean that you’re really talking to a server room instead of a cute brunette. These are just some of the latest headlines that made me question whether the term Ethical AI is increasingly becoming an oxymoron.

If you’re not concerned, I genuinely envy you, or maybe you’re just less cynical than me and believe in the high road and believe in things like compliance, and governance, and the biggest word of all: legislation, which is moving at glacial speed in this country, but realistically in most other countries.

But since this is a marketing and media publication, I shall try and get your attention with something closer to home, or in this case our daily jobs in the marketing profession. How are you currently making sure that your ads, or your clients’ ads for that matter, are not used to train third party providers’ LLM (large language models) and literally infringe your copyrights?

There are multiple Australian start ups currently pilfering ad libraries from Google and Facebook pulling advertisers’ ads into their tools, to literally replicate them by AI.

According to MagicBrief’s own ads, it spells out that you can “steal” with AI, and once you scroll to the T’s and C’s they make it clear that they believe it’s only a copyright infringement if advertisers inform them.

So right now, third parties are capitalising on your investment in marketing expertise, your own ingredients for successful marketing, your style of imagery, your copy and your brand.  Just one of the many, many things no one seems to be paying attention to.

I believe this is because many marketers are completely unaware of the risks posed by a new breed of AI services, aside from all ethical considerations.

On a more human level, my concern is that mindless automation will lead to marketing that is detached from empathy. For example, if we start using AI to create hyper-personalised marketing messages, (Publicis kicks off 2024 with one-to-one wishes, powered by AI)  human intervention is essential to determine whether customers or staff  will feel like their privacy has been violated - but I guess that again would require stronger privacy laws, which are currently elusive but looming with incoming new legislation.

If you’re concerned after reading this, make sure you get your (or your clients’ assets) out of any third party tools ASAP.

Or maybe ignorance is indeed bliss. Highly efficient AI bliss.

See Also: Jasmin Bedir: The trap of shiny new AI tools
See Also: Jasmin Bedir: Limitations of AI in an advertising agency

–

Top Image: Jasmin Bedir