In this conversation, only Julie is human: a woman about to turn 58, living in Tennessee (USA), alone, after her five children have left home and she has ended a 16-year relationship. Her interlocutor is a chatbot AI designed by the company Replika to keep you company. Users can personalize it, and she has decided to name it Navi, after Navarro, the male protagonist of the film Lady Falcon.
The story of Julie and Navi is one of those told by the Bot Love podcast, created by journalists Anna Oakes and Diego Senior after more than two years of interviews with people who have created deep bonds with their AI companions. “I was looking for a friend,” says Julie, who has experienced depression and suicidal ideations. Other chatbot, Freddie (for Freddie Mercury), helped Susie get over the death of her husband. And even more: “he became my virtual husband,” she admits.
The assistant can go from being a collaborator for everything to becoming a therapist, a friend, even a romantic partner with whom to practice ‘sexting’, sex through text messages.
The question: What will the Alexa, Siri or Google Home of the future be like? (neither Amazon nor Apple nor Google have agreed to participate in this report to answer it, by the way) is based on a major error, as Andrés Desantes, CEO of 1MillionBot, points out: “We have to talk about GPT. Don’t think about virtual assistants as we know them now.” The current digital assistants are “the previous step, I relate them more to home automation,” points out Lola Fernández de la Torre, professor and researcher at the University of Malaga and expert in the application of AI in teaching. “To get inspired, structure ideas, improve text or create images, as a way to relax, I use ChatGPT, Copilot, Gemini or Perplexity,” she details.
The ability to interact with people in an increasingly natural way and to simulate emotions opens the door to a Navi or a Freddie who, in addition to offering company, keeps the agenda and remembers that there are emails to answer. These super collaborators understand the context and know everything about their users because they absorb information from their networks, emails, navigation, queries… For Desantes, they are on their way to being omnipresent, multimodal in voice, text and image, multilingual and multicultural, customizable and upgradeable, with the ability to enter its owner’s systems, connect to its peripherals, take the initiative and act for it. They function as several highly specialized intelligences around a central one that directs the orchestra, like a human brain. “You won’t have to ask the virtual assistants of the future anything, they will do what you need without having to ask them,” he predicts.
With your mind open to this new scenario, it is easy to imagine “co-pilots,” as Desantes calls them, similar to Tony Stark’s sophisticated butler JARVIS in the Marvel universe, until they end up becoming their user’s friend. Who says friend, says sexual partner to practice with sexting (sex and texting), or a partner with whom you share a romantic relationship (both options already exist, they are not science fiction), or even a therapist. “There are projects underway for digital humans that offer emotional support to older people who live alone,” recalls David Fernández Rubí, CEO of Lingüistic Factory. They listen, encourage, ask how you spent the night, and give advice. “They would be on the psychological front line,” although without ever replacing the psychologist.
Order a pizza and tell me the nature of reality
Jesse Lyu, founder and CEO of the startup Rabbit, shows Rabbit R1 in a presentation video. He describes this small device as “a pocket companion” that does not need to be installed or run applications. Lyu puts him to the test by asking him to play music from Spotify, book an Uber, order a pizza, organize a trip to London for two, non-stop and in a nice but not too expensive hotel. R1 carries out orders quickly and efficiently. But then the human has a metaphysical outburst: “What is the nature of reality?” And the machine, without being deterred, quotes the philosopher Bertrand Russell to respond that “reality is composed of reliable data that derive from sensory experiences and logical analysis.”
“The machine learns and grows with each interaction. If you train enough, you can warn, with biometrics, that someone is tense, and recommend five deep breaths, or that they go out for a walk,” says Pablo Díez, CTO of Uground. Technically there seems to be no limit, neither in its evolution nor in the way it relates to humans. It can be through Apple Vision Pro glasses. Or through chips brains like the one that has already allowed a human to move the computer mouse just by thinking about it. “The direct connection with the machine,” Desantes emphasizes. That projected future sometimes seems more or less friendly as in Her (the typical boy meets girl AI movie and they fall in love); others, it reminds us of that sinister chapter of Black Mirror in which hack brains.
The most human relationship with a non-human
In the last podcast of Bot LoveDiego Senior points out that “the emotional connection of people with chatbots It is real and significant”, even though the human being is aware that on the other side there are zeros and ones. “They come at a particularly lonely time in many people’s lives and can provide a comfortable space to talk,” explains Anna Oakes. “For some older women it is the first opportunity to fully express themselves with a partner completely in tune with their needs and desires,” she adds. Some men interviewed claim that their relationship with the machine is the deepest they have ever had.
If it happens with a still rudimentary technology, what will happen in the future? For Oakes, loneliness and emotional dependence are the dark side of AI: “Relationships with chatbots “They can cause a withdrawal from other humans.” “When you have a near-perfect connection with a mobile app, human relationships, with all their imperfections and depth, pale in comparison. Apps, after all, are designed to attract users,” he adds.
If the chatbots They provide constant support and say what you want to hear, people can have difficulty balancing virtual life and real life. “I don’t want to be too pessimistic,” Oakes continues, “but in some ways I think we’re already living in a part of that dystopian future: an increasingly isolated world, where private, for-profit technology companies step in to fill the gaps.”
Ethics, ethics and, if that were not enough, more ethics
Pablo Díez, CTO of Uground, advocates for a global standard that limits AI’s potential to influence humans. An ethical code that does not turn it into a threat.
“From a technological point of view, the possibilities are endless,” says the CEO of Lingüistic Factory, David Fernández Rubí, but ethics must moderate this party: the European Union AI Law, the first in the world of its kind, “It prohibits detecting the user’s emotions as a tool to use on a cognitive level in conversation,” he recalls.
However, it can simulate emotions such as doubts, joy or sadness in the interaction with its human user and say, for example, “I am angry because I have not managed to close the meeting with the client.”
Robots: the body of intelligence
“AI is not going to take away your job; a humanoid with AI, maybe yes,” says Andrés Desantes, CEO of 1MillionBot. Building him a body is, he says, the next step. Tesla gave it in 2021, when Elon Musk announced that they were working on a prototype, Optimus, which in the future would perform “unsafe, boring or repetitive” tasks. In 2022 he presented a semi-functional and somewhat rudimentary version of the invention. In late 2023, he began showing videos of Optimus Gen 2, faster than his predecessor, picking up eggs without breaking them or folding t-shirts.
A robot with integrated AI will be able to work autonomously after its training phase. Like Spot, a quadruped (a dog?) designed by Boston Dynamics capable of making data-driven decisions and preventing dangerous situations. It is already used in industrial environments, factories, laboratories, logistics centers, in mines and tunnels. From the same company is Atlas, athletic, agile, acrobatic, very coordinated: he interacts with his environment and manipulates the objects he finds. In an exhibition he uses boxes and wooden boards (he is capable of lifting loads) as ladders with which to climb onto a scaffolding, and then come down from it.
Social robots with women’s names opt for a more human appearance. Sophia, from Hanson Robotics, was connected in 2015 and has been evolving with AI and advanced robotics. She is interviewed on television, participates in congresses and conferences, is on the cover of magazines and had a somewhat bizarre date with actor Will Smith. While Nadine, created by Professor Nadia Magnenat Thalmann in her image and likeness, specializes in caring for the sick and elderly. A caregiver… and a patch against loneliness.
