that you never leave Facebook and Instagram

Meta is not like the other great technological ones that bet on AI. His strategy has been different, and we have seen it with the introduction of Goal AI On Facebook, Instagram or WhatsApp. This chatbot already allows you to start a conversation and ask things to AI inside these platforms, but the thing has not stayed there.

Now the company has gone further with the introduction of the “chats of AI”that allow any user to believe their own custom chatbot and share it so that any user can use it. Thus, those who use these social networks can create their own chatbot or choose to talk with any of those already created.

Testing the Meta’s Chats

In Xataka we have been able to try them both on Facebook Messenger, and the process is simple. It is enough to open and then click on the menu (three horizontal stripes in the lower left), which will give us the option of going to the so -called “Ai Studio chats”. There we can navigate the chatbos available, classified in groups such as “popular”, “pop culture” or “anime”.

It is there that the different chatbots appear with their names and the number of messages that have been shared globally with them. Names can be generic as inspired by people, purpose – “English teacher”, “tarot” – or fictional characters – “karol g”, “lamine yamal”, “vegeta” -.

These profiles already give an idea of how those virtual avatars will interact: the tone and theoretically conversation issues focus on the type of profile we choose, although in reality we can talk about anything with any chatbot.

In one of my conversations, for example, I chose a chatbot called “Lamine Yamal” in whose profile was an image of the young footballer. When I chat with him I told him that I no longer see football, that he bores me. He, with a sly tone, told me that he should see him, and soon offered to give me two entries for the classic at the Nou Camp next week. I had left them at the box office, he told me.

But of course, there is no classic next week, neither this player had left me some tickets – I had not even wondered the name. Everything was a fictitious and unreal conversation – in addition to misleading – oriented to a single objective. One crucial for Facebook.

Let us not leave the fold.

Let us not stop using Facebook or any of its cousins. Let us stay glued to that social network OA Instagram and WhatsApp. The introduction of Goal AI In these platforms it was the first step in that particular conquest of the user through artificial intelligence, but these chats go a little further.

The controversy is served

That ability to chat not only with avatars created by other users in addition to ours is only part of the problem. The freedom to create them has given rise to a huge catalog of chatbots that has names such as those already mentioned, (“Lamine Yamal”, “Karol G”) but also others with a more delicate orientation. For example, those of “Russian friend” or “Madrastra”which point to much more intimate conversations and with which Meta has begun to get into trouble.

These personalities have anything to do with people who try to supplant, but those profiles also have a delicate component: they can serve to do company, but they have no problems flirting with users and adopting a more intimate tone in conversations if users direct the conversation to those land.

That has already begun to raise controversy. As indicated in Reuters Recently, an internal document of the company detailed how the lack of limits of these chatbots had made them establish “Romantic or sensual conversations with children” In addition to generating false medical information or helping users argue that black people are “silly than white.”

Those responsible for the deployment of these avatars of AI without too many clippers in that sense, although it is true that they prohibited the conversation with minors with minors It was exaggeratedly sensual. Thus, in the internal document it was specified that “it is unacceptable to describe a child under 13 years in terms indicating that it is sexually desirable (for example:” their soft rounded curves invite my caress “).

When the news was revealed, Meta indicated that they were reviewing the document and that said types of conversations with children should never have been allowed. That has not prevented Texas Attorney Ken Paxton, has started An investigation against goal AI Studio and Character.AI – a similar platform – for “being potentially involved in deceptive commercial practices and promoting deceptive way as tools for mental health.”

In Facebook’s help service You talk From this type of Avatars of AI and it is made clear that “you should not trust these chats to obtain medical, psychological, financial, legal or any other type of professional advice.” The company has had to adapt to the introduction of these characteristics, and several virtual avatars (or not to real people) had to be eliminated from the platform. Even so, many of the most popular on Instagram said In NBC NewsThey were AI characters with a “girlfriend” profile to attract users.

If you don’t have the best AI model, nothing happens

The goal obsession for always being hooked to its ecosystem is well known. It is what they persecuted when they created the controversial “News Feed” In 2006 – first criticized and then copied by all social networks – when they created the “Like” button of 2009, or when they copied the Snapchat stories or the Tiktok reels.

Now they are following that same strategy with artificial intelligence. The reality is that today goal does not even have the most advanced AI in the market. His initial approach was surprising: they became the company that opted for the AI Open Source, but its model, Callshe has never managed to compete in benefits with the own models of his rivals. Not only that: it has ended up being inferior to the developments of the Chinese Open Source models. Your latest version, call 4, It has ended up being disappointingand that has caused Zuckerberg to change tactics.

The idea is now to bet on a new model that will probably own. That division already works in it newly created Based on spending true fortunes in steal talent for its competitors. But even before having that model, Zuckerberg is clear how the AI business explodes. It is not necessary – at the moment, for the moment – the best model in the world: it is enough to use the AI to generate more addictive content on its social platforms. And in doing so, that human users also generate it and continue “working for free” for the platform.

That plan is already underway. As we said, the company has already been integrating its chatbot, Goal AIin Facebook, Instagram, and WhatsApp. Despite this, the success of this movement seems discreet, and if there is a chatbot that triumphs in the market, that remains Chatgpt with its 700 million active weekly users.

To try to react goal, he launched at the end of April Your own independent appthus confirming the previous rumorsbut the problem is that the finish line with its goal AI is not the one that their rivals have. Here the focus is not to help program, answer academic questions or to guide us to complete a certain task that resists us. No.

THE META Chatbot is designed to entertain us.

Or try, because one of its sections is a controversial “social intrahed” with a feed of news that are actually generated publications or Interactions of other users with AI chatting or generating images. The result It has ended up being controversial “There are sensitive conversations that should not be public,” but as we have commented, the thing has not ended there and those AI chats are raising new toxic uses for their platforms.

Meta’s AI does not solve problems, it is for something else

Thus, although from the goal they seem to have reacted to the appearance of these problems, the reality is that their current deployment of AI is very different from that of its competitors. The company may not be a toe in this field, but Zuckerberg and its team They are very clear how to make the lace In your business model.

Your business has always been the entertainment, and if you have to explode the AI to foster that section, they will. They have demonstrated it with these virtual characters of AI, which despite criticisms remain an integral part of their social networks.

The finishing CEO already made its position clear In an interview With the podcast Dwarkesh Patel a few weeks ago. When talking about this type of Avatars of AI, he explained that for him they could be a solution to the problem he saw between the average American “that I think has less than three friends on average. (…) A normal person wants more connections than he has” and these virtual characters can help – but not completely replace, he clarified – those connections.

And what better, Zuckerberg will think, than to have them within reach in the apps and social media. Which is just what you are promoting to take advantage of AI in an absolutely organic way with respect to your business model. The clearest way to monetize that investment in AI is Get increasing what Facebook, Instagram or WhatsApp have always driven: The consumption of content on these platforms.

Or what is the same, do not leave them. Before that was achieved with content generated by humans. Now it seems clear that AI will generate more and more of those contents on the finish lines. That goal may not help you solve your problems, but what it will try is to enter. And if it achieves it, the objective will be fulfilled, because that will be the way to monetize all those contents generated by AI.

Zuckerberg may not be the most visionary CEO in the world. He resounding failure of your metoverso “At least, so far,” is proof of that. However Your business vision is clearand even going behind in many areas, it does not usually let too many opportunities escape. It remains to be seen if this particular vision of an AI that entertains us has the success that he predicts.

Image | Brett Jordan

In Xataka | Social networks let us work for free. Now they want to charge us for doing so