Artificial intelligence raises a true revolution in our world, but for now its impact is relatively reduced. The appearance of Chatgpt In November 2022, things began to change, and although the rhythm of advances is remarkable, the scope of AI is limited.
In fact, we have been talking about three large types of systems of systems for a long time artificial intelligence. Let’s see them with some more detail.
So or artificial superintelligence
Beyond the general artificial intelligence (AGI) is the call Artificial superintelligence (like)that would be that capable of creating systems that exceed human intelligence in all metrics.
Thus, one would be among other things capable of improving itself autonomously, refining and optimizing its algorithms exponentially. That capacity, together with its cognitive superiority, would allow such a global challenges such as climate change, the shortage of resources or potential pandemics, and do it with an unparalleled effectiveness.
But the development of a also raises a threat important potential. If these systems end up solving problems that go beyond human understanding, thus could redefine all kinds of industries, revolutionize the industry and change our society of unpredictable ways.
In addition to how such could become aware of itself, thus being able to develop their own desires, motivations and moral framework, what would make it unpredictable.
It is precisely that possibility that is reflected in many dystopic science fiction films such as’2001: an odyssey of space‘ either ‘Terminator‘in which the human race ends up practically extinguished by the machines. And it is precisely one of the risks Of those who speak Some experts In ia. Others like Yann Lecun, head of the goal, see that potential threat as “Absurdly ridiculous”.
Agi or general artificial intelligence
Much more ambitious than weak artificial intelligence is General Artificial Intelligence (AGI)which would allow solving any intellectual task to resolve by a human being. This artificial intelligence would be multitasking and could do hundreds, thousands of different things well.
A general artificial intelligence would be able to perform all the tasks performed by human beings and even others that are not capable.
Over time it is estimated that this type of agi systems could replace the human being in virtually any field And they could make human labor obsolete, something that would have gigantic social and economic implications.
These systems could replicate the human ability to reason, learn and adapt to new problems.
Thus, theoretically an AGI would be able to generalize their knowledge and apply them to a wide number of scenarios, something that the weak or narrow systems cannot do, although modern models of reasoning with Openai O1/O3 or Deepseek R1 go in that direction .
Ani or artificial intelligence “narrow”
These types of systems are able to solve very well defined and limited problems. The Narrow artificial intelligence (ANI) It is the one that has caused the true explosion of this discipline in recent times: different techniques have been applied such as Deep machine or learning To solve specific problems, and the results have been exceptional.
The artificial vision systems that apply for example in urban traffic management are a good example of ANI.
This artificial intelligence, also called “weak” or “narrow” (ANI, by artificial Narrow Intelligence) is what we have long seen applied to many scenarios with greater or lesser success.
The achievements achieved with Deep Blue or with Alphago They are a perfect example of weak artificial intelligence: they solve a specific and delimited problem and that allow them to be solved so that these systems end up performing those tasks much better than a human being.
But also now they are much more present in the hands of All generative AI models They are available in services such as Chatgpt, Claude or Gemini, but also in others such as Midjourney or Dall-E 3. These models are not able to adapt to their environment, although they can maintain a coherent dialogue, as the aforementioned chatbots do.
In Xataka | The best books to enter artificial intelligence: five experts reveal their basic readings
GIPHY App Key not set. Please check settings