 |
| Wallet scam Spanish film "Los Tramposos" |
About 30 years ago, I received my first attempt at the Nigerian scam: a letter from someone unknown to me offering me the chance to participate in a capital evasion operation involving several million dollars, which they proposed depositing into my bank account, offering to share the profits in return. Of course, I didn't reply, although I kept the Nigerian stamp that came with the letter. I suppose that if I had replied, they would have asked for access to my account to make the transfer, and instead, it would have been emptied.
Since then, I've received dozens of similar attempts, since a certain time via email, always originating from some African country. This scam is, in a way, similar to the wallet scam, because in addition to the swindler, the victim also tries to commit fraud; in one case against the supposed mentally challenged person offering the alleged money, in the other against the governments affected by the capital evasion.
As we entered the 21st century, new forms of fraud
emerged, such as phishing,
a cybercrime technique that impersonates companies (usually banks) or trusted
individuals, to steal confidential information, passwords, or banking details.
Deceptive messages, via email or SMS, are used to direct victims to fake
websites that mimic the originals, thus stealing their data. I've received many
of these myself.
With the rise of new "artificial
intelligence" tools (language
models, image generators, voice generation programs, "AI" agents), even more sophisticated forms of attempted fraud
are emerging.
Although I don't participate in the most widespread
social networks, such as Facebook or Twitter (aka X), I do use others, like LinkedIn and Goodreads
(the social network for books). On the latter, I was the target of a new type
of attempted fraud that took me some time to detect.
Around September 2025, I received a request to
become "friend" of a Goodreads profile I didn't recognize. It had
just given a good rating to one of my books, as if it had been read and enjoyed.
The “person” behind the profile claimed to be an author, and indeed, several
books were listed in that profile. At that time, Goodreads had a rule that two
different profiles could only communicate if they were "friends."
I accepted the "friendship" request and
immediately received a message praising my book, offering to discuss my work,
and asking me questions. I investigated and discovered that there were two
profiles on Goodreads with the same name, one of which was older, and the other
(the one that had contacted me) newly created. This struck me as odd. At one
point in the conversation, the profile in question said in a message: Chatting
through Goodreads is complicated. Why don't we stick to WhatsApp or Telegram? Obviously, someone wanted my phone number, and I
suspect they would eventually offer me author services (book promotion,
facilitating film or television adaptations, etc.) for a price I would never
recoup.
Reviewing the messages I received, I noticed their
writing style closely resembled that of language models. This made me
suspicious, and my suspicions grew when the profile in question disappeared
from Goodreads a few days later. This process, using different real author
names, was repeated in the same way at least four times over the following
months. At a certain point, I stopped accepting the "friendship"
offers and reported the case to Goodreads, who then deleted the duplicate
profile.
I suspect this has happened to other authors as
well, because at the end of 2025, Goodreads announced that starting in January,
they would no longer allow message exchange, even between "friends."
Since this change took effect, the aforementioned attempts have ceased.
Another type of attempted fraud has recently been
reported on LinkedIn. Through this social network, a person received a message
from a headhunter offering a job with a very high salary. Everything seemed
very professional, but the affected person became suspicious when he saw that his
questions were answered instantly, at any time of day or night, and on
weekends. He also noticed that the conversational style of his interlocutor
resembled that of language models. Finally, upon discovering that the LinkedIn
profile in question was newly created and lacked a history (as in my case on
Goodreads), he concluded that he was communicating with an “AI” agent. This is his
conclusion:
This is
going to get worse. And we all know it. Much worse. The tools to set something
like this cost less than €20 a month. You don't need a team. You don't need a
company… And most people won't even notice.
This video describes a similar situation, carried out via a phone call, with a hilarious ending. Less hilarious is the fact that technological advances can facilitate the commission of certain crimes, such as identity theft. And what about this recent article in IEEE Spectrum? As Noam Holt says in this article, So my thought was not just ‘this is disturbing,’ but also ‘what’s next?’
Thematic Thread about Natural and Artificial Intelligence: Previous Next
Manuel Alfonseca
No comments:
Post a Comment