If you’ve lived in Romania long enough, you know the stereotypes. Bank transfers to and from the country are a headache.
It’s even worse if you have Belgian or Portuguese citizenship. Countries that give citizenship to people who just live there for five years end up giving citizenship without vetting them to well. This means that being a Portuguese or Belgian citizen is no guarantee of predictable behavior patterns. How can you tell if someone is Portuguese when anyone could be Portuguese? It is almost like having no passport at all.
Being unable to tell the real from the fake is a problem. Here is a comment on one of my posts that might have been created by AI:
It’s fascinating to see how Romania’s supermarkets are evolving with self-checkout systems and improved shopping experiences. The changes seem to focus on convenience, but I wonder if they’re equally accessible for all customers, especially older or less tech-savvy individuals. The expansion of bagging areas at Lidl is a great improvement, but I’m curious about how these changes impact the overall shopping speed and efficiency. Do you think the shift to larger baskets is practical for everyone, or does it exclude those who prefer smaller, handheld options? Also, how do store assistants manage the balance between assisting customers and maintaining the self-checkout flow? Overall, these updates seem positive, but I’d love to hear more about the challenges and feedback from shoppers.
Ten years ago, spammers did not bother actually talking about my post. They would just say vague things like “Well written” or “you forgot to check your facts” or mention my SEO or something like that. They would use a bot to leave the same comment on five thousand different blogs. So, it was easy to see that no one read my post because the comment would have nothing to do with my post. A post on Lidl in Cluj in 2017 wouldn’t have spam mentioning “Romania’s supermarkets” or “the self-checkout flow.” So, I would know the comment was genuine.
However, after this, there seems to be an advert for a product. The AI can summarize a post, and create questions on it. It’s not very hard to train a bot to “read a post, create an AI generated response to it, adding an advert, and submit that as a comment.”
That said, there are hints within the post itself that it was either generated by AI, or by someone who didn’t read my original post. It asks a question that we already answered, that in fact was one of our main complaints in the original post.
And there are other clues. It is written in an academic way, but there are people who do talk like that. Not a lot of people use terms like “the self-checkout flow” but it happens. If I spend a lot of time looking at academic journals, or on LinkedIn, I might do it myself.
Anyway, if we look at Google’s new terms of service, we see that those in Europe have an option to opt-out of much of the spying. Most companies, like Facebook, only allow you to fully opt out if you get the paid version. Well, I am not sure what the exact writing will be in the new Google terms yet, but it seems that you need to find a special hidden button and press it to assert your European rights. (Non-Europeans don’t have these rights.)
While it is illegal to generate a fraudulent AI image of a celebrity to sell an investment scam without labelling it AI, it doesn’t seem to be illegal to create AI generated comments. (Although spam itself breaks certain rules.) Nor does there seem to be much regulation on labelling comments as created by AI. Sadly, this means we might be rejecting some comments because of the sheer volume of garbage we get.
Most of our research is done in person, through live conversations, and not in front of a computer. This will have to increase as we find email and social media updates, as well as comments on our posts, and even news stories from major sources, to be less reliable than before. Which is a good thing, it’s always nice to go out for a walk and meet real people, see real things, and so on.