What is Artificial Intelligence?

A brief history of AI

Despite its current relevance and modern characteristics, artificial intelligence actually traces its roots back to the early 1900s, when the idea of robots and artificial humans was sparking the curiosity of inventors.

The concept of the robot was first introduced and popularized by Czech playwright Karel Čapek in 1921, and by the 1940s the idea of artificial brains and smart machines became a major area of interest. It was in the 1950s that Alan Turing proposed the Turing Test, a criterion for measuring machine intelligence that laid the conceptual framework for modern AI.

In the years that followed, AI research saw rapid progress. The 1980s marked the AI boom, which brought major advancements, including the creation of expert systems and the first attempts at autonomous vehicles. Despite a short setback due to diminished funding and interest, now known as the AI Winter, the 1990s saw a turning point for AI with milestones such as IBM’s Deep Blue defeating world chess champion Garry Kasparov. During this time artificial intelligence made its way into everyday life, with products like speech recognition software and Roomba - a robotic vacuum cleaner.

The 2010s marked an explosion of new possibilities and a new era of rapid AI growth. Breakthroughs such as self-driving cars, AI-powered assistants like Alexa or Siri, and language models and chatbots have embedded AI into many of our daily routines.

Where is AI headed now?

Current AI trends are progressing towards more specialized and efficient models in generative AI, as well as towards agentic artificial intelligence, in which separate models join forces to complete user requests faster. As these trends continue, we accelerate towards a future with autonomous vehicles, AI-generated multimedia, AI assistants in our daily lives and other life-changing developments.

AGI and ASI: the future of machine intelligence

What we commonly refer to as AI can be broken down into three levels of artificial intelligence: artificial narrow intelligence (ANI), artificial general intelligence (AGI), and artificial superintelligence (ASI).

Artificial Narrow Intelligence is what we are familiar with today; AI systems designed to perform specific tasks with a limited scope of abilities. Unlike human intelligence, ANI is highly specialized and excels at one function, such as image recognition, language translation, or playing chess, but lacks general problem-solving ability.

Artificial General Intelligence is the next big goal in AI development. AGI is an AI system that would function on a similar level as human intelligence, being able to understand, learn, and apply its knowledge across a wide range of tasks. AGI would be able to think critically, adapt to new situations, and exhibit reasoning and common sense.

Artificial Superintelligence is the hypothetical stage where AI surpasses human intelligence in all aspects, including creativity, problem-solving, decision-making, and emotional intelligence. An ASI system would not just be smarter than humans—it would think in ways beyond human comprehension.

How would AGI differ from today’s AI?

Unlike current AI, which relies on massive datasets for training, AGI would learn in real time like a human (though probably faster and more efficiently). While AI excels at specific tasks, AGI would apply intelligence across multiple fields. Additionally, AGI would have a true understanding of context, emotions, and abstract concepts, going beyond merely processing data.

Despite AI advancements, no AGI currently exists. The biggest challenges in achieving it include common sense and adaptability, as AI struggles to generalize knowledge across different domains. Cognitive reasoning is another hurdle, as machines lack true self-awareness and understanding. Additionally, ethical and safety concerns must be addressed, as AGI would require strict guidelines to ensure its safe deployment.

Many researchers believe AGI could emerge in the next few decades, while the more optimistic predict it in the coming year. Currently, the timeline remains uncertain.

The potential and risks of ASI

The potential benefits of ASI include superhuman problem-solving, enabling it to tackle complex issues in physics, medicine, and engineering. It could also self-improve exponentially, leading to an intelligence explosion, and display advanced creativity, generating groundbreaking scientific discoveries, art, and literature beyond human ability.

While ASI holds the potential to revolutionize technology and human progress, it also presents significant risks. These include existential threats if ASI is not aligned with human values, the potential loss of human control as ASI may make decisions beyond our understanding, and ethical dilemmas about who should control an intelligence greater than humanity itself.

Given that ASI is purely speculative at this point, many researchers emphasize the importance of AI safety measures now, before the development of AGI, to ensure a secure foundation for further development.

Where are we now?

We are firmly in the AI stage, where machines assist humans in specialized tasks. AGI is still theoretical, with ongoing research aiming to bridge the gap. ASI remains completely hypothetical, though its potential raises critical questions about the future of AI.

As we advance toward more powerful AI systems, understanding these distinctions will help society navigate the challenges and opportunities of artificial intelligence.