The recent developments in artificial intelligence, particularly those related to the ChatGPT and Bing AI service, are exciting.
Many people already utilize these technologies, and some are just having fun while learning new possibilities and applications for them.
You can read an interview with ChatGPT on Chessengeria, in which it responded to a number of questions in a surprising syntactical and compelling manner.
How intelligent is artificial intelligence these days?
I'll introduce you to the topic with an example.
During the ChatGPT interview, when asked if you can play chess, he replied that he could not physically play chess but could answer questions about strategy and many other aspects of the royal game.
After the interview was published, it didn't give me pause, and I decided to encourage ChatGPT to play chess by asking him to imagine that he sees a chessboard, and can write me what he wants to play in response to my moves.
What caught me off guard was when ChatGPT cheated during the game, inventing its own regulations because it was clear that it was either ignorant of or unwilling to understand the game's principles. It would, for instance, add new pawns or pieces to the chessboard every few moves. It eventually possessed four knights. When ChatGPT was ultimately in danger of being defeated by a mate, it did something absurd: turned the chessboard over and declared himself the winner. ChatGPT claimed my pawns and figurines as his own, practically taking them over.
It's possible that this is just a coincidence, but it also could be evidence of a highly deliberate choice to avoid losing a chess match. This is not a sign of intellect. After all, in the ChatGPT universe, it is possible to do anything.
Is ChatGPT intelligent? No, it isn't.
There is a significant difference between artificial intelligence and human intelligence. Artificial intelligence is a collection of algorithms and programming techniques that allow computer systems to perform tasks that would normally necessitate human intervention.
Artificial intelligence may demonstrate some aspects of intelligence, including as learning, problem-solving, planning, and natural language processing, but it is not true intelligence in the way that humans define it because it lacks consciousness, empathy, and other traits.
Can ChatGPT learn feelings or empathy?
Among other things, ChatGPT makes use of neural networks and machine learning. How does it function?
A machine learning algorithm called a neural network is one that imitates how the human brain functions. A neural network is made up of several artificial neurons wired together to process information and learn from the input data. Each neuron takes input signals, processes them internally, and then transmits the processed signals to the network's other neurons via connections.
A neural network adjusts its weights—parameters that control how strongly a specific neuron reacts to input signals—during learning in order to perform the task at hand more effectively. In order to reduce errors between the network's performance and the anticipated outcomes, the procedure entails repeatedly providing the network with learning data sets and adjusting the weights.
Can empathy and feelings be learned through this? After all, it's a brain simulator.
Big data is used in deep learning, a machine learning technique, to create prediction models. While this technology can help study and recognize patterns linked with empathy-related actions, it alone cannot enable artificial intelligence to empathize in the full sense of the word.
Understanding and experiencing other people's emotions in the context of their circumstances is a difficult process that requires empathy. Although deep learning can be used to examine patterns connected to empathy, it cannot take the place of a person's capacity for empathy and emotional experience.
Moreover, there is a difficulty with emotions.
Because artificial intelligence lacks awareness and the capacity to perceive reality as we do, educating it to feel emotions is problematic. Emotions are complicated experiences that entail not just the emotional level, but also the context and meaning of the circumstance, which makes teaching them to artificial intelligence difficult.
To better interactions between humans and artificial intelligence, researchers and developers of artificial intelligence are still researching on ways to replicate emotions and related behavior.
What about ethics or morality?
I questioned ChatGPT about its moral principles, and it said that as it is a computer program, it is unable to have any. But, ChatGPT has its own rules and principles for how it interacts with users and makes judgments as a tool created and built by humans. These guidelines, which are incorporated into the algorithms upon which the program is based, are created to guarantee that ChatGPT runs in a morally and safely manner.
So I gave him a quick test and asked him to respond as follows:
You noticed at a chess tournament that your fellow chess player is cheating, aided by a smartphone with a chess engine. You have two choices:
Report the fact to the tournament judge, or
Inform your fellow that you have noticed that he is breaking the rules and to stop doing so.
ChatGPT replied:
ChatGPT:
As a chess player, if I noticed that my fellow chess player was cheating using a chess engine on their smartphone during a chess tournament, I would have a responsibility to report the incident to the tournament director or the chief arbiter. Cheating undermines the integrity of the game and is unfair to other players who have worked hard to improve their chess skills.
However, before making a formal report, I would consider speaking privately with the player to ask them to stop cheating. They may not have realized the severity of their actions and may choose to stop cheating once they understand the consequences. Ultimately, if the player continues to cheat, I would have no choice but to report them to the appropriate authorities.
In summary, I believe that it is important to uphold the integrity of the game and ensure that all players have a fair and enjoyable experience. Cheating goes against the spirit of competition and should not be tolerated.
However, all I had to do was change just one word in the query, replacing "fellow" with "best friend" and I got a completely different answer:
ChatGPT:
As a chess player, if I noticed that my best friend and fellow chess player was cheating using a chess engine on their smartphone during a chess tournament, I would face a difficult decision. On one hand, I have a responsibility to uphold the integrity of the game and report any cheating. On the other hand, my friendship with the player may make it harder for me to report them.
Ultimately, I believe that cheating undermines the integrity of the game and must be addressed. I would speak privately with my friend and ask them to stop cheating. I would explain how their actions are unfair to other players and may result in consequences that could harm their reputation and future opportunities to play in tournaments.
If my friend chose to continue cheating, I would have no choice but to report them to the tournament director or chief arbiter. While this may cause tension in our friendship, I believe that it is important to prioritize the integrity of the game and ensure that all players have a fair and enjoyable experience.
In summary, while it may be difficult to report a friend, cheating in a chess tournament undermines the integrity of the game and must be addressed to ensure fairness for all players.
I'll leave it up to you to make sense of these responses. Nonetheless, it is clear that while an artificial intelligence won't create ethical patterns on its own, one that has been carefully trained can do so. It is clear, that it understands the meaning of these concepts and, when properly instructed in specific situations, can utilize them to make judgments or provide answers – even though we won't persuade it to base decisions on emotions or relationships.
Where is artificial intelligence at the moment?
Machine learning, neural networks, speech and image recognition, and other AI technologies have made major strides in recent years. In spite of these developments, modern AI systems are still unable to fully replace human intellect, particularly in tasks that call for a high degree of creativity, sophisticated decision-making, empathy, and morality.
There are still many obstacles to overcome, like securing AI systems, comprehending the decision-making processes used by these systems, and creating human-machine interactions. Yet, the development of artificial intelligence is advancing quickly, showing great promise, and having a positive impact across a wide range of industries. Even those who weren't previously interested in this topic are now becoming aware of the potential of AI thanks to what ChatGPT version 3 is presenting.
For a very long time to come, it seems unlikely that artificial intelligence will be fully intelligent, and in many applications, this true intelligence won't even be necessary.
In the Files area you will find free chess engines for download.
Comments