The use of artificial intelligence has skyrocketed on all fronts. While the focus is on ChatGPT and the algorithms that generate images, the reality is that the integration of the technology into workflows goes beyond a simple chatbot. A case that has gained strength in recent days is the publication of articles written by IA in CNET.
At the beginning of the month, Futurism reported that the technology website has published dozens of articles generated with artificial intelligence. Automation goes beyond one or two paragraphs, as it CNET he gave the green light for them to be written in their entirety. There are more than 70 pieces signed under the pseudonym CNET Moneyall focused on financial advice and presumably verified by the editorial team.
The use of artificial intelligence to write full articles has generated controversy among the public and some members of the press. Far from the fear of being replaced by an AI in the short term, the criticism focuses on the lack of transparency of the medium with its readers. Each post features a disclaimer stating “This article was assisted by an AI engine and reviewed, verified, and edited by our editorial team,” albeit in such small typeface that it would go unnoticed by the general public.
The veracity of the content was also questioned. A second article from Futurism he discovered that some articles have blunders in basic finance concepts. The artificial intelligence that writes in CNET misinterprets the information, resulting in miscalculations of interest on a loan or earnings from investing.
Soon you will not be able to tell if the article was generated by a human or an artificial intelligence
Given the controversy, CNET defended the use of automated tools in writing articles. Connie Guglielmo, editor-in-chief, published a piece that try to make posts transparent with artificial intelligence from your finance section. Guglielmo mentions that they discovered errors such as incomplete company names, transposed numbers or confusing language.
A point that draws attention is the reference to the plagiarism. In accordance with The Verge, 41 of 77 posts were corrected and now include the warning: “We replaced sentences that were not entirely original.” Guglielmo revealed that the editor could not detect sentences or phrases that resembled the original source. “We are developing additional ways to mark exact or similar matches with other published content identified by the AI tool,” she said.
AI engines, like humans, make mistakes.
Connie Guglielmo, editor of CNET
CNET He was clear in his position and assured that they will continue to explore the use of artificial intelligence in their editorial workflow. The outlet warns that media companies and content creators need to learn (and understand) about the use of automated tools. Although your application is for minor content, this does not guarantee that in the future we will not be able to distinguish if the content was written by a human or by an AI.
AI vs. content creators
The use of artificial intelligence as a tool to automate creative work has created friction on several fronts. On the one hand we have the ChatGPT case and its prohibition in some schools in the United States. As teachers seek to adapt the educational model to take advantage of technology, some students from prestigious universities —like Stanford— confessed that they cheated by using ChatGPT in their final exams.
There is also the content issue with copyright. In order to function properlyAI engines have to feed on content and learn from it. Tools like Midjourney and Stability AI are already the focus of a copyright infringement lawsuit. A group of artists accused the companies of training their artificial intelligence models with images without consent.
The fight against generative AI has been fought most vigorously in this arena. For months, artists have been holding protests in communities such as artstation publishing images with the “NO AI” logo, forcing the web to impose filters and labels on automated content. The lawyers for Ortiz and company consider the images generated by artificial intelligence as stolen content.
While the lawsuits will be resolved and the legislation will adjust to the new changes, it is delusional to think of artificial intelligence as a passing fad. The explosion of ChatGPT and DALL-E are the beginning of a new era in which, like it or not, automated tools will be part of our day to day. “We adapted to calculators and changed the tests in math class. This is a more extreme version of that, for sure.” Said Sam Altman, CEO of Open AI.