Carlos Núñez, president and CEO of PRISA Media, emphasized that LLMs or large language models have been training for a long time and that it is important for media groups that their content is not being used to develop new business models. "The AI Act requires transparency and traceability to see what data has been used so that copyright can do its job," he highlighted. Núñez noted that "there is a need for data to feed these models," which is why some companies are becoming "more active and aware" and willing to enter into agreements, seeing "the present and future use of their data" as beneficial.
Evelio Acevedo, Managing Director of the Museo Nacional Thyssen-Bornemisza, also participated in the panel. In his opinion, the digitization or technologization of art "is neither new nor simple." He pointed out that the first AI creations in the art world were already happening in the 1960s. Although there were technologies like the printing press that facilitated the work of writers, artists, etc., the difference now is "the loss of control." "From a museum's perspective, we cannot forget how and for what purpose AI-generated content is used. We must remember that we work for the people," he stated. For Acevedo, "an AI-generated painting can evoke emotions if the artist has conveyed it well. But for artificial intelligence, there are things difficult to capture, like emotions... They are the result of a series of spontaneous reactions. So, an AI-generated work might aesthetically appeal to us, but there is something much deeper," he defended.
Ángel Durández, president of Arcadia Motion Pictures, the studio behind Blancanieves and Robot Dreams, echoed this sentiment. Although Robot Dreams deals with the relationship between an anthropomorphic dog and an AI (a robot), the process of this silent animation film could not be more artisanal. "We had a team of 80 artists, and we did everything manually. We managed to do something completely outside the current trend, away from AI," the studio head boasted.
TV presenter Luján Argüelles pointed out the fundamental difference that "a machine will not be able to observe and notice nuances as a human does." She argued that when making television, 60 people are involved in building an "almost perfect machine." Additionally, she believes that "you must find where the magic is in your content, and AI will never have that. The magic in a book... a screen will never achieve that. Culture must be touched and enjoyed, and that always happens by being more human," she maintains. Argüelles insisted that many things need to be regulated and organized because the industry focuses on launching products, and "problems arise later."
Opportunities of Generative AI
Despite these reservations, the panel participants did not deny that generative AI presents many opportunities for the audiovisual and artistic world. The head of Arcadia highlighted that it has reduced the costs of series and movies. He mentioned how some cities have started charging for photographing their spots to avoid the expenses of traveling, meals, etc., for film crews. "These cities want a small piece of the pie." Additionally, he revealed that in Blancanieves, they digitally included the bull alongside the lead actress. However, for him, the key is that "new generations understand what is real and what is fake."
For the president of PRISA Media, there are many opportunities in terms of monetization and productivity, but "an AI will never replace journalists because you have to find and chase the news."
Argüelles commented that AI forces us to be good journalists because currently, "there are thousands of news stories that are nothing, just seeking clicks." She invited reflection on "the type of journalism we want to do." Nonetheless, the presenter defended that "we need to know where things come from. The media is one of the sectors that need the most regulation" and argued that "viewers have the right to know that what they hear is true and not manipulated by AI." Acevedo also argued that "the media should be accountable when something is false or misleading" and that critical, analytical thinking should be encouraged, helping people understand what everything is.
Are Data Running Out?
Núñez raised a potential issue if large language models continue to 'feed' at this rate because "everything on the internet is already trained." According to the head of PRISA Media, data are finite and could run out. This has led to models being trained with synthetic data, which results in more hallucinations from AIs.
The panel also touched on the topic of NFTs, which became a trend a couple of years ago and generated significant speculative movement. For the museum director, they have helped democratize art collecting. "We understand that when we were kids, we collected cards, and life evolves," he compared. Regarding artificial intelligence, Acevedo explained that it allows them to save a lot in production and work processes and helps them "know our audiences very well." "As for rights, we have no choice but to partner with big tech companies, thinking about common knowledge. We must create the largest knowledge database," he defended.
In the case of cinema, the head of Arcadia did not want to speculate on whether a film generated entirely by AI would ever win an Oscar, but he assured that examples of movies about artificial intelligence like 2001 or Blade Runner show that "reality has far surpassed imagination."
What AI Will Never Replicate
Carme Artigas concluded the panel with a question for reflection: "Is there any inherently human trait that AI will never replicate?" Acevedo responded that while generative artificial intelligence can be creative, its creativity will never be based on emotions and sensations; it will not be "emotionally deep creativity." Durández agreed. In his opinion, feelings like love or fear are "hard to replicate" with AI. "We are going to reinvent ourselves almost as a species," he predicts.
For Luján Argüelles, humans are "a product of a miracle," which makes them unique. Lastly, for Núñez, what 'a machine' will never be able to replicate is "the entire moral and ethical dimension of our decisions."
