ChatbotGPT: more marketing than a threat to journalism
The journalistic environment began to experience a kind of technological frenzy after the release of the ChatbotGPT program, seen by many professionals as a threat to the exercise of the activity and the consequent increase in unemployment among reporters, editors and commentators. The reality is that these fears are only partially true because artificial intelligence will mainly affect low-skilled professionals who follow the “copy and paste” technique.
The technological infrastructure of the new digital fad is neither new nor unprecedented. It emerged more or less 80 years ago with the generic name of Generative Artificial Intelligence (Generative AI), that is, a digital program capable of generating innovative texts, images, sounds and codes. In this category are programs like ChatbotGPT and others like Bard, produced by Google.
The great differential of software such as GPT and Bard is the ability to work with an immense, almost immeasurable volume of data digitized on the internet that serves as raw material for sophisticated robots/algorithms to perform combinations and recombinations of facts, events and numbers. The GPT is, for the time being, a quantitative advance in the digital information datafication process. The answers from Generative AI programs tend to become more and more complete as more data is digitized.
The emergence of new types of chatbots is inevitable, whose differential will be in the size of the database on which they will operate. There already exist, for example, software, such as Dall E, which produces images from keywords provided by a person. We are witnessing the beginning of a race for the market of chatbots included in the umbrella of Generative Artificial Intelligence. For now, it is a quantitative process, with more data and better answers, but it is likely that with the arrival of quantum computing, we will see something qualitatively new in terms of intelligent machines.
The analysis of texts produced in various journalistic experiences with the Chatbot launched by the company OpenAI, linked to Microsoft, shows that it is good at assembling texts commonly known in newsrooms as “archives” or background stories. They can save reporters and editors time when they summarize past facts and events. By sifting through the avalanche of data published on the internet, the GPT gathers the most mentioned elements seeking to produce minimally balanced and diversified final versions, when the object is a controversial topic.
Artificial intelligence will affect the performance of journalists because it will automate the production of texts and supporting images, giving experienced and qualified professionals more conditions to produce more creative content with greater analytical depth. This will require greater training and knowledge from reporters, editors, commentators, illustrators and programmers since less complex tasks will be able to be performed by intelligent robots capable of sifting through zillions of digitized data and facts on the internet in an infinitely faster time than that of a human being.
Psychology of algorithms
The problem is that the answers given by Generative AI programs need to be reinterpreted by journalism professionals. Our current language is full of ambiguities, biases, and subjectivities that the machine can interpret incorrectly and give answers based on this false reading. The problem is not with the software, but with who made the query or request. So, for the answer to be credible, the journalist needs to have basic programming knowledge to minimize the mismatch between what the professional wants and what the program thinks the reporter or editor wants, as explained by Professor Nicholas Diakopoulos, from Northwestern University in the United States in an interview with Columbia Journalism Review.
It is then obvious that the great challenge posed by artificial intelligence to the exercise of journalism is that of professional qualification not only in content production but also in the relationship with new digital technologies. Writing well is no longer the main virtue of a good journalist, but his ability to expose data, facts, events and ideas using multimedia resources allows the greatest possible communicability and interactivity with the common citizen.
Programs such as ChatbotGTP create the need for new skills and competencies in the exercise of journalism, such as the relationship with digital technologies. It’s more than knowing how it works and how to use it. The professional will need to understand and interpret the machine, in a kind of psychology of algorithms (interview in Portuguese), as defined by neuroscientist Álvaro Machado Dias. It’s just that the robots will have to be taught to collect what data we want as well as understand how they will search so that the results can be considered adequate to the objectives we are looking for. An algorithm tutor in an artificial intelligence system knows that the universe of searches has become immeasurable in the face of the volume of accumulated data. Therefore, the results may be unpredictable given the incalculable number of possible correlations between the archived data.
Data Colonialism
The transcendental importance of databases is also clear. They are what make artificial intelligence possible. For this reason, the British researcher Nick Couldry, from the London School of Economics, warns about the emergence of the so-called “data colonialism”, a phenomenon that can alter the relations between nations and companies. In the new digital economy, those who have data have everything, which can turn world politics upside down, since companies like Alphabet (owner of Google and Youtube) and Meta (owner of Facebook and WhatsApp) now have more information about the world we live in than any global superpower like the United States, China, and Russia.
The already unimaginable amount of data held by companies like Facebook, Google and Twitter end up conditioning our way of living, thinking and acting because changes in the programming of the algorithms that govern interactivity within social networks can alter, unexpectedly and unpredictably, the way we do business, live in communities or make our political choices. Database owners increasingly own the world, and GPT is the latest weapon of these cyber colonists
