Artificial intelligence does not rebel.. Who is really responsible?

The writer began her article with a humorous anecdote during her summer vacation, when she discovered that one of her nephews, an elementary school child, was able to connect to the hotel's Wi-Fi using a password he had learned during a previous stay. She asked sarcastically whether the child had "gone out of control" or reached a level of "super-intelligence," before concluding that he is simply a child who is good at pattern recognition and has a strong motivation to reach his goal.

Navaroli used this incident to clarify her point of view regarding what AI companies describe as a rebellion of their systems, emphasizing that unexpected behavior does not mean the machine possesses independent will or human consciousness.

Language that hides responsibility

The writer believes that attributing human qualities to computers is an old practice that predates the emergence of large language models. She cited the incident where OpenAI launched a chatbot with a voice that sounded very similar to the voice of actress Scarlett Johansson in the 2013 film "Her," before the company pulled that voice.

Navaroli says that presenting systems as human entities makes the public view their outputs and errors as actions taken by a machine with a will, rather than holding the entities that designed them and determined their operating mechanisms accountable.

The problem, according to the writer, is exacerbated when journalists adopt the terminology used by technology companies to describe their products. For this reason, the Associated Press warned journalists against portraying artificial intelligence systems as entities that think, feel, want, or understand, and called on them to focus on how they work, their efficiency, the entities that developed them, and the people who might be affected by them.

Linguistic alternatives and reliable sources

Navaroli recommends that journalists consult with statistics, computer science, and data experts to accurately understand technical terms, and refer to critical research that addresses the limitations of artificial intelligence and how it works.

She also calls for replacing vocabulary that suggests the machine possesses human qualities with more precise expressions; instead of talking about the "answers" of chatbots, the word "outputs" can be used, which helps describe what the systems produce without implying that they think or understand in the human way.

The writer cites the position of researcher Emily Bender, a professor at the University of Washington, who emphasizes that extraordinary claims require extraordinary evidence, pointing to the need to deal cautiously with exaggerations regarding the capabilities of artificial intelligence and its alleged consciousness.

Amplifying fears serves companies

Navaroli asserts that the task of journalists is not to create hype around artificial intelligence or echo the warnings issued by companies about the future of humanity, but to provide accurate coverage based on evidence that clarifies the actual capabilities and limitations of this technology.

She believes that portraying systems as more conscious and independent than the evidence proves may enhance the market value of technology companies and help create new wealth as some of these companies prepare to go public. It may also give them greater influence in the debate regarding the regulation of the technology, by entrenching the idea that they are the most capable of understanding and controlling their systems.

At the end of her article, Navaroli revealed that she had used Google's AI-generated summaries to find some links, but said she has avoided using these tools since she began publicly disclosing her use of them, due to feeling embarrassed about having to admit it.

The writer concludes with a striking irony: transparency in the use of artificial intelligence may, at times, lead journalists to refrain from using it altogether.