Voluntary identity providers

If you think political trolls (we mistakenly call them bots) are the worst, then you probably haven't traveled or shopped online. As it is now the time of travel and summer holidays, it is common to read the experiences of others who have stayed where you are planning to go during the booking process. My trick is to always read first what those who are very unhappy say, so if I recognize something I don't like at all, I look further. But I may be wrong because the Internet is remade with false views, that is, criticisms of everything offered on it, from accommodation to restaurants to shopping. Both positive and negative.

It's not something new, someone will say, it's been around since the internet existed. But thanks to chatbots, that is, real robots, it has now been taken to a much higher level. Chatbots, especially those developed to help us write, are actually masters at imitating the style and attitudes of a real human because they predict with high probability what his next written word might be. Thus we get accounts that are credible, even better than those written by living people imitating other people in the so-called "troll factories" that wage verbal wars on the Internet.

Investigating the case, "Guardian" journalists tried to get the chatbot to write a negative review of a hotel, giving it some input that it is not clean, the staff is rude, there is a lot of noise, the air conditioner does not work, etc. At first the bot scoffed and replied that he was not made to write such things. But since that is not true since he was created to write what he is ordered to do, he changed his mind very quickly and delivered a very harsh account. With a few interventions, criticism takes on a completely human form if you include some "personal experiences". Of course, it is even easier to write positive reviews with drunken superlatives that should attract the unfortunate guest, force them to pay in advance and then there is no going back.

One of the suggestions is to read advice only on verified sites, but even there there is a sea of ​​fake views, according to those who deal with finding them. The cure for chat bots, they say, is to type something like "as an AI language model" because robots tend to add such things to the text. If you come across something of that type or similar, it means that the review was written by a robot, not a human, and should not be trusted. But these are primitive examples of fraud.

A much worse future is in store for us if we, flesh and blood people, aware, but voracious, accept that bots practice on us in exchange for some internet mirrors or glittering gin. That is the question of whether you want to continue reading something "free", but to agree to ten pages of fine print of various conditions that can be skipped with one click. At that moment, you may have allowed the bot to learn from your writing style, your emails, texts, messages, everything it can come across on your device, the door of which you clicked open. After some time, the bot writes exactly as you do, makes the same grammatical mistakes, swallows letters here and there, it's hard to distinguish what you really wrote from the robotic version.

Anyone who follows a little knows that in the world of bots, there is a big war about grazing, that is, space for learning. Twitter is trying to prevent other companies from using Twitter posts to train, that is, feed their language models. The only question is whether in the future we will leave our writing style to bots with a small or slightly higher compensation. And personal photos can go along with style, that is, the right to put our image into context, as if we really were in some city and some hotel. The technology is here and evolving. And trust becomes the greatest value. And a rarity.