Artificial Intelligence (AI) is becoming a common tool in everyday life. People trust it with routine tasks and use it in complex technological processes. Nevertheless, AI remains an assistant that plays a secondary role.
The so-called Artificial General Intelligence (AGI) can change the rules of the game. In science fiction, it is often a machine that can think and realize itself. That is, an "artificial intelligence" that is as intelligent as a human being.
In the modern world, AGI is often perceived as an inevitable stage of AI development. To understand whether this is true, we should answer a few questions.
What is AGI?
The first of these concerns what general artificial intelligence is in the minds of researchers, businesses, and society at large. In the expert community, AGI is often referred to as a system with human or superhuman mental abilities.
CEO of DevRain, AI enthusiast, and author of ChatGPT, DALL-E, Midjourney: How Generative Artificial Intelligence is Changing the World, Oleksandr Krakovetskyi, in his commentary for Mezha.Media, explains that AGI is formally an artificial intelligence capable of transferring knowledge and experience from one area to another, performing a wide range of tasks at the level of human intelligence or higher.
He adds that while this definition seems clear, the specific technological criteria are still unclear. In this context, he mentions ChatGPT, which is trained on textual data but is capable of conducting market research, writing code, and analyzing business models.
"Is this AGI? Most experts agree that it is not. Modern models are powerful, but their internal structure is still far from true general intelligence," emphasizes Oleksandr Krakovetskyi.
When will AGI appear?
Another important question concerns the timing of the emergence of general artificial intelligence. Is it realistic to create it with the current state of technology? As with the definition of AGI, there is no clear answer here either.
In his essay, Dario Amodei wrote that many people are skeptical about the emergence of "powerful AI" in the near future. According to him, some people do not believe that it will ever be created.
"I think it could happen in 2026, although there are options where it could take much longer," the Anthropic CEO predicted.
According to him, most human knowledge is not reduced to text alone, but is based on our perception of the world. What people take for granted is difficult for computers. That is why Ian LeCun is convinced that human-level AI "will require profound changes in perception."
Oleksandr Krakovetskyi points out that AGI was previously expected to appear no earlier than 2040. But now more and more experts are changing the estimate to 2030. There are suggestions that something very similar to general artificial intelligence may appear even earlier. To explain this, he cites several reasons:
- Impressive pace of development: previously, new breakthroughs in AI occurred every few years, now they are almost monthly or weekly.
- Unprecedented competition between companies and governments: OpenAI, DeepMind, Anthropic, xAI, and other players are investing billions of dollars in the development of artificial intelligence, which accelerates the race.
- A breakthrough is possible through new architectures: if researchers find a way to overcome the fundamental limitations of modern neural networks (for example, to teach models to understand cause-and-effect relationships), progress could be even faster.
"At the same time, there is a possibility that at some stage we will face fundamental problems, which may suspend the race for some time," DevRain CEO adds.
Potentially, one of the problems could be the lack of data for training language models. According to OpenAI co-founder Ilya Sutskever, who left the company to found Safe Superintelligence Inc.
What are the threats from AGI?
Experts have different assessments of the technology's future. Alexander Krakovetsky, for example, believes that two main scenarios are possible in the development of AGI.
According to one of them, general artificial intelligence will be created before it is officially recognized. The expert notes that this may be a deliberate strategy to gradually adapt society and markets to a new level of technology, reducing panic and regular pressure.
"AGI will not be created, but another improved system will be called AGI. This can be used as a way to reduce public fears about real threats or to stimulate investment in technology or for marketing and hype," he says of another possible scenario.
However, speaking about AGI, experts also focus on threats from general artificial intelligence. DevRain CEO emphasizes that the biggest one is humanity's unpreparedness for such a rapid development of technologies that will have a large-scale impact on various areas, including:
- Labor market - many professions may disappear or change significantly, causing a wave of social upheaval;
- Business companies will have to adapt quickly or give way to competitors who will implement AGI more effectively;
- Education - traditional teaching methods may become obsolete, and qualification requirements will change dramatically;
- Policy - governments will have to quickly develop a regulatory framework to control the development of AGI and prevent its abuse;
- Society - AGI can exacerbate inequality issues if control over the technology is concentrated in the hands of a few corporations or states.
"We should also not rule out possible existential threats if AGI gets out of control or is used for military purposes, although for countries like ours this is a chance to defend ourselves against the backdrop of the failure of international institutions and the insignificance of international guarantees," emphasizes Oleksandr Krakovetskyi.
Many Western experts are also concerned about serious threats. They warn of potential cyberattacks, the development of new types of biological weapons, and other risks. Therefore, they call for technology control to avoid any pessimistic scenarios.
What opportunities will AGI offer?
Meanwhile, general artificial intelligence is not only about threats but also about great opportunities for humanity. Oleksandr Krakovetskyi cites a breakthrough in medicine as one of the enormous benefits, as the technology can improve diagnostics and help find a cure for incurable diseases.
Another significant advantage is the automation of routine tasks, which will allow you to perform them faster, cheaper and with fewer errors.
In addition, AGI can become a personal assistant to help with learning, information analysis, and decision-making, thereby significantly expanding human capabilities.
Other benefits include optimization of economics and logistics. The company's CEO is confident that AGI can radically improve production, transportation, and energy consumption.
However, with the advent of general artificial intelligence, other opportunities will open up. If people treat technology responsibly, we will see interesting changes in many areas.