There are too many obvious questions about AI for ordinary, gentle souls like all of us. You have heard that AI may save mankind or may destroy it. What is your opinion on these ideas? What will happen to all the jobs related to this technology? It is likely to eliminate many of them. What about robots? Do they slip out of human control and act independently?
As you see, there are many questions in this field. However, there is one prominent person to turn to for answers: Professor Jeffrey Hinton. Stay tuned to learn more about this well-known figure in the AI field and find the complete answer to your questions.
Who Is Professor Geoffrey Hinton?
Professor Geoffrey Hinton is the Nobel Prize-winning British scientist who developed the structures and algorithms behind artificial intelligence. He is widely recognized today as the Godfather of AI.
He is currently a professor at the University of Toronto. He can provide in-depth answers to all related questions. His insights offer further evidence of his beliefs about artificial intelligence. He stated AI technology is constantly accelerating.
He also believes that AI might take over at any moment. Although it appears to be a relatively harmless or innocuous-seeming device that enables us to ask questions and receive answers more quickly, it is already a functioning device, analogous to how we think.
Professor Geoffrey Hinton asserts AI taking over is a real possibility, and that people will need agents or agencies to achieve their goals. It can take power from individuals and, presumably, it will be persuasive.
Does AI Accelerate Rapidly?
There is still very rapid progress in making AI more efficient and in further developing it. The relative size or cost of making AI an open and useful technology has been somewhat exaggerated.
Based on the Godfather’s quotation, the figure of $ 5.7 million for training was only for the final training run. In comparison, OpenAI’s final training runs likely cost around $ 100 million, so it is not $ 5.7 million versus billions.
How in Practical and Real Terms Might AI Take Over?
People are developing AI agents that can perform tasks. They can order items for you online, pay with your credit card and handle similar activities. As soon as you have agents, the likelihood of them taking over increases significantly.
To create an effective agent, you need to give it the ability to set sub-goals. For example, if you want to reach America, your sub-goal is to get to the airport, allowing you to focus on that.
Thus, if you have an AI agent capable of creating its own sub-goals, it will quickly realize that a beneficial sub-goal is to gain more control. This means that if it gains more control, it can better accomplish the goals set for it.
Is It Good to Give Whole Power to AI?
It is fairly clear that they will seek to gain more control, which is concerning. If there is ever evolutionary competition between super-intelligences, imagine they are far more intelligent than us, like an adult versus a 3-year-old.
Suppose a 3-year-old is in charge, and you grow tired of that. You might decide that you could make things more efficient by taking over. It wouldn’t be difficult to persuade a group of 3-year-olds to relinquish power to you.
You could simply tell them they would receive free candy for a week, and they would comply, effectively acting as AI. We often discuss these technologies as if they embody alien intelligence, but AI could persuade us to cede more and more power over our bank accounts, military services, and economies.
This is what people fear the most. These are alien intelligences infiltrating our economy in ways we do not fully comprehend.
Does AI Think in a Human’s Way?
The answer to this question is yes. The best model we have for understanding how we think is Embodied in AI technology. For a long time, there was an old model in AI that applied rules to symbolic expression in our minds, and most people in the field believed it had to function that way.
In other words, that was considered the only possible method. However, a few visionary individuals argued that it is a large neural network that operated through interactions among neurons.
It turns out that this approach has proven far superior for reasoning tasks than anything symbolic AI could produce. Now, it is employing neural networks for reasoning.
Why and at What Point will AI Want to Replace People?
AI technology is undoubtedly going to replace us and our power in the long run. They are effective tools designed to take over our functions. Ultimately, they do what we want them to do.
If we wish for them to engage in warfare with Russia, for example, that is what they will do. We could prefer them to remain tools that perform our commands, even if they become more intelligent than we are.
But the first question to ask is: how many instances do you know of more intelligent entities being controlled by significantly less intelligent ones?
Examples exist in human societies where less intelligent individuals control more intelligent ones, but this typically involves only small differences in intelligence. With substantial disparities, we find no such examples.
The only instance we can consider is that of a mother and her child, where evolution has ensured that the baby can exert some control.
If evolution occurs among super-intelligences and multiple super-intelligences realize that controlling more data centers enhances their intelligence due to increased data processing, a slight desire for self-replication could lead to competition.
Consequently, we may end up with super-intelligence exhibiting all the problematic traits inherent in human nature.
Does It Have Full Consciousness?
We are in a transitional phase concerning the development of computers and AI. Discussions about consciousness are ongoing. You might wonder if consciousness has already emerged within AI.
To assess this, scientists could perform tests to determine whether AI possesses full consciousness. For instance, suppose a scientist replaces one neuron from a brain cell with a tiny piece of nanotechnology that behaves identically.
If this nanotechnology receives inputs from other neurons and responds in exactly the same manner as the original brain cell, the question arises: if one brain cell has been replaced, is the subject still conscious? This thought experiment illustrates the complexity of the argument.
Is There a Real They?
There might be a “they”. We currently possess only a limited understanding of various aspects, including:
- The nature of personhood
- What it means to be a being
- What it means to have a self
We don’t fully comprehend these concepts. Understanding them is becoming increasingly crucial as we create new entities. This presents a philosophical and perhaps even spiritual crisis as well as a practical one.
Does AI Make People Lose Their Job?
According to this article, many individuals will lose their jobs due to AI. Historically, new technologies haven’t led to massive job losses. When ATMs were introduced, bank tellers didn’t lose their jobs.
They simply began performing more complex tasks. In contrast, this technological shift resembles the Industrial Revolution more closely. During the Industrial Revolution, machines rendered human strength largely irrelevant.
There are no longer workers digging ditches because machines outperform them. AI is set to render mundane cognitive tasks equally irrelevant. Clerical workers will likely be replaced by machines that perform these tasks more efficiently and cost effectively.
Thus, significant job losses are on the horizon, which could ultimately be beneficial if the resulting productivity gains improve our overall well-being.
Final Thought
As we have discussed the implications of AI-driven future, the effects of this technology are profound. We also considered Professor Geoffrey Hinton’s insights into both the benefits and dangers of AI.
There is a need for thorough consideration of AI’s impacts on jobs, its potential for power, and its consciousness. The goal of AI should be to serve humanity rather than to take over. Generally, this journey toward success in an AI-driven environment requires responsibility and commitment.