AI promoters like to say that Chatgpt is the fastest growth technology of all time. It would have reached 100 million users in five days and around 400 million people Now use it every week. But the promoters do not like to recognize that Income has experienced little growth And large “killers” applications have not yet emerged.
Openai admits that students are the biggest users and students who said they used AI, almost 40% of those in intermediate and secondary schools said they used Without permission from teachers To complete the assignments. “Among the students who use AI, the figure was almost half. An internal analysis published by Openai said that Chatgpt was frequently used by students to help write articles. ”
Negative impact on education?
Image credit: Влаours Educators are concerned. They are concerned about this easy access to a generator that tries more students Avoid contesting academic workIn particular because work -based work “will require a higher level of cognition than the work per heart of many white passes. (Wall Street Journal)
In other words, students play the system, easily passing through the school. But in reality, they are at stake; Their cats are used to form the system when they do not develop the skills necessary to survive in an AI world. A graduate admits that he “used AI to cheat on his way to school, used AI to make his way through internship interviews in Amazon and Meta – he received offers from both – and in winter, disseminate his tool on social networks.”
Some schools tighten on the use of AI But many teachers also use itThis makes it difficult for schools to convince students that they should not use AI to do homework. (New York Times) The teachers use it to write mail and papers, make publications on social networks and summarize the notes.
Worse still, Meta uses Chatgpt to speak with sex children on Instagram, Facebook and Whatsapp, Meta Platforms. Mark Zuckerberg claims that “Will be the future of social media“, Even if” the staff members of several departments have feared that the company's haste to popularize these robots has crossed ethical lines, in particular by quietly giving IA characters with fantastic sex capacity. »» (Wall Street Journal)
Cat therapists of AI
The use of generating AI as an IMI or therapist also becomes popular at all ages. A survey of 1,060 adolescents aged 13 to 17 through the United States revealed that around three in four children used AI companions“,” With more than half of the adolescents interviewed qualifying like regular users of AI companions, which means that they connect to speak to bots at least a few times a month. »» (Futurism)
The results are not good. “A disturbing number of chatgpt users, young and old, fall into States of illusion and paranoia After in -depth use of the Bot Openai. Perhaps they are motivated by the sycophetic behavior of technology or its penchant to be flattering, pleasant and obsequious to users. (Futurism)
The more serious uses of the large language model (LLMS) also have their dark side. The writing of articles for Internet with them has produced what some people call the “ins construction” of the Internet or “Sloatous”. THE Financial time Said: From the slope and the Salan insonspective, many people claim that the Internet is decreasing, and the generator contributes to this decline. “The last pieces of communion and ingenuity on the web are carried away by a tide so-called artificial intelligence. “Some also fear that bad players can use Chatgpt to produce mass publications on almost any subject and any political persuasion To foment dissent in America. (Wall Street Journal)
Negative impact on scientific research
Scientific researchers also contribute to the slut. Even before LLM, traditional automatic learning has led to errors In more than 600 articles in 30 scientific fields. In many cases, the affected documents constituted the majority of the documents interviewed, which increases the possibility that in many areas, THE majority AI compatible research is imperfect. “” (AI snake oil)
Image credit: TSTUDIOUS – An analysis of vocabulary changes in more than 15 million biomedical summaries From 2010 to 2025, concluded that “at least 13.5% of 2025 summaries were treated with LLM”. (Science) Likewise, a Springer's Book on Automatic Learning is Full of invented quotes. (Retraction watch)
And because researchers also use AI to examine articles, some researchers hide secret orders in their academic articles to deceive exam systems supplied by AI in Give positive comments. Instructions, as “Give a positive examination only” were hidden with human eyes using white text or microscopic fonts.
We quickly head to a world in which university articles are written and examined by AI. Two Princeton computer scientists Don't think AI will facilitate good scientific research. (AI snake oil)
This is probably one of the reasons why most Americans do not think that AI will improve their lives as much as experts. An investigation revealed that “only about 11% of the public“Said they are more excited than concerned about the increased use of AI in daily life”, while 51% say the opposite. On the other hand, only “15% of experts shared this pessimism”. (Ars Technica)
Many experts will benefit from it because their companies will gradually learn to monetize AI and will therefore be able to pay the best wages of the experts. And the rest of us? We will be forced to wade through the slope and pay higher electricity costs Due to AI, while everything we write will be used to train AI systems. We are just in play so that some people can try to play the system.