Вештачка интелигенција

Artificial Intelligence: Debunking the Myths

Summary

Artificial intelligence (AI) has become a topic of great interest among technology leaders, politicians, and the public. While some enthusiastically promote technologies like ChatGPT as a valuable tool with the potential to transform society, others express concerns that any tool […]

Artificial Intelligence: Debunking the Myths

Artificial intelligence (AI) has become a topic of great interest among technology leaders, politicians, and the public. While some enthusiastically promote technologies like ChatGPT as a valuable tool with the potential to transform society, others express concerns that any tool labeled as “intelligent” has the ability to surpass human capabilities.

Professor of Philosophy and Psychology at the University of Cincinnati, Anthony Chemero, argues that the confusion surrounding AI stems from linguistic misconceptions. He asserts that while AI is indeed intelligent, it cannot be intelligent in the same way as humans, even though it can “lie and talk nonsense like its creator.”

According to Chemero, our everyday usage of the term AI is certainly intelligent, but intelligent computers have existed for many years. In a co-authored paper published in “Nature Human Behaviour,” he explains that ChatGPT and other AI systems are large language models (LLMs) trained on vast amounts of data collected from the internet, much of which contains the biases of the individuals posting that data.

“LLMs generate impressive text but often invent things out of thin air,” Chemero says. “They learn to produce grammatically correct sentences, but they require far more training than humans. They actually don’t know what they’re saying means,” he adds. “LLMs differ from human thinking because they lack a physical body.”

While those who create LLM models may call it “hallucination” when they invent things, Chemero suggests it would be more accurate to call it “making shit up.” LLMs explicitly add a word that is statistically likely to be the next, but they don’t know or care whether what they say is true.

With a little prompting, he asserts that it is possible to make AI tools say “ugly things that are racist, sexist, and other forms of bias.”

The aim of Chemero’s work is to emphasize that LLM models are not intelligent in the same way humans are because they lack the capacity to care. “Things matter to us. We are dedicated to our survival. We care about the world we live in,” he says, noting that LLMs don’t exist in the world and are indifferent to everything.

FAQ:

1. What is artificial intelligence (AI)?
Artificial intelligence is a field of computer science that focuses on developing computer systems capable of performing tasks that typically require human intelligence.

2. What are language models?
Language models are computer programs that use statistical methods to predict the most likely next step in a sequence of words based on previous data.

3. What are the benefits of artificial intelligence?
Artificial intelligence can bring numerous benefits, such as automating complex tasks, faster and more precise data analysis, process optimization, and improved resource management.

4. How can artificial intelligence have biases?
Artificial intelligence can have biases if it is trained on data that contains biases from humans. For example, if models are trained on texts that contain sexist or racist attitudes, there is a risk that such attitudes will be reproduced in the generated texts.

Source: University of Cincinnati