AI experts say we are on the wrong way to achieving human AI

According to a group of hundreds of researchers of artificial intelligence, the field is currently pursuing artificial general intelligence in the wrong way.

This insight was revealed in the Association for the Development of the Presidential Committee of Artificial Intelligence (AAAI) for the future of AI Research. Thehe long report They were collected by 24 AI researchers whose experience varied from the state of AI infrastructure to the social aspects of artificial intelligence.

The report included a basic takeover of each section, as well as a section of public opinion, where respondents were asked their own thoughts about the section.

The “AI Perception of Reality”, chaired by MIT’s computer scientist Rodney Brooks, refers to the characteristic of the Gartner Hype cycle, a five -step cycle commonly found for a technological super. In November 2024, Gartner “estimated that he had just crossed his peak and was at a decline over,” the report said. 79% of the respondents in the Public Opinion section said the present public perception of AI capabilities does not correspond to the reality of the AI ​​Research and Development, with 90% saying that the non -compliance interferes with the AI ​​Research – 74% of that issue, saying that “AI studies are guided by AI.

Artificial General Intelligence (AGI) It refers to human intelligence: the hypothetical intelligence of a machine that interprets information and learns from it, as a person would do. AGI is a sacred grab of the field, with consequences for automation and efficiency in countless areas and disciplines. Consider any tribal task that you do not want to spend a lot of time from traveling to your taxes. AGI can be deployed to ease the weight of a company tasks, but also catalyze progress in other areas, from transport to education and technology.

The surprising majority – 76% of the 475 respondents – have indicated that just a scanning of current approaches to AI will not be enough to get AGI.

“In general, the answers show a cautious but still moving approach: AI researchers give priority to safety, ethical governance, sharing benefits and gradual innovations, advocating joint and responsible development, not race to AGI,” the report said.

Although the overwhelming state of research – the current approaches to the AI ​​do not put the researchers on the most optimal path to AGI – technology has made jumps and borders.

“Five Years Ago, We Could Hardly Have Been Having This Conversation – Ai Was Limited to Applications Were a High Percentage of Errors Could be Tolerated, Such As Product Roommend Stictly Circumscribed, Such as Classifying Scientific Images, “Explained Henry Kautz, A Computer Scientist at the University of Virginia and Chair of the Report’s Section on Factual & TRust “Then, quite suddenly in historically, General Ai started working and reaches public attention through chatbots like Chatgpt.”

AI Facturality is “away from resolved,” the report read, and the best LLMS answered only about half of a set of questions in a comparison test in 2024. But new teaching methods can improve the stability of these models and new ways to organize AI can improve their performance.

“I believe that the next stage of improving reliability will be the replacement of individual agents of AI with cooperating teams of agents who are constantly checking the facts or trying to preserve each other honest,” Kautz added. “The bigger part of the general public, as well as the scientific community – including the community of AI researchers – pursues the quality of the best AI systems today; the perception of AI lags for about a year or two behind the technology.”

AI does not go anywhere; After all, the Gartner Hype cycle does not end with “fading in oblivion”, but instead the “Plateau of Performance”. Different arenas of AI uses have different levels of super, but with the entire AI clam for AI – from the private sector of government officials, the hell, our own families – the report is a refreshing reminder that AI researchers think very critically about the state of their area. From the way AI systems are built to the ways in which they are introduced into the world, there is room for innovation and improvement. As we do not return to time without AI, the only direction is ahead.