No, my point is that you cannot teach a machine morality any more than you can teach it emotion. Any detail brought up or contrived during this dialogue is a mere commentary on that fundamental fact. Absolutely. It is a necessity. No, a machine has no ability to "disagree". That's the point ..... it doesn't because it can't.
There's that word "moral" again. Morality doesn't make the rules. It is only a human construct used to support the action of those who sway power and influence. That's not always true. The perceived harm (to the group) can be seen as avoided by the murder of individual group members. But again it is by the way of calling murder something else and is effectively a cure for "a bad apple in the barrel" and can be considered morally justified.
What state do you live in? Denial? I asked AI what is the best way to save 20 out of many and it came back with Random Selection is the fairest way. That is a moral judgment. It avoids the pitfalls of your preferred method of using personal bias to judge the relative value between individuals. If you put in a false statement into AI it will disagree with you. If you tell it to do something immoral (how to build a bomb) it has algorithms to guard against that and says no. It is still in its early stages and learning, like a child, right from wrong.
It has two names: Reality and Honest endeavor. If you are quoting it verbatim then I believe that's what it said. No, it most certainly is not. To believe that "random selection" is a moral judgment makes me wonder what state you live in. Denial? Personal bias is not being discussed here. You have incorrectly introduced it on your own initiative and it is false. It cannot "disagree". It cannot distinguish "morality". You're intentionally not paying attention. When you write such things do you not realize how silly it is?
When confronted with your scenario, rather than being told to choose the first 20, I asked AI what is the best way to choose the 20 that would survive and it said: “Random selection: To ensure fairness and avoid favoritism”. That’s an example of an ethical decision decided upon through an algorithmic process. It is a fair and a moral one, I would also say a higher ideal than yours which attempts to gauge the value of individuals which is highly subject and prone to unavoidable bias. I could choose the 20 youngest or the 20 smartest; would that be the most “fair”? You've clearly have not worked with AI....why don't you ask it some difficult questions involving morality..... https://chat.openai.com/
As I've said all along, I take your word for the results you've been given. The problem is that a "random selection" is most definitely NOT the best way to survive. Common sense will tell you that random selection is a bad idea. The very fact that you mention, "To ensure fairness and avoid favoritism" means that either the programming is flawed or that perhaps you've entered stipulations that require the machine to give a contradictory reply. To put it simply (and this ought to be clear without needing much explanation) "survival rate" and "fairness/favouritism" are at opposite ends. The goal is either survival or social fairness/political correctness. You cannot have both. The success of one denies the success of the other. You can call it what you like but the results you've been given are not due to a moral-learned mechanical decision. If you have decided that you want a "fair and moral" outcome then that is translated into your own programming but the "fair and moral" aspects are not weighed by the machine in any way at all. You call it "fair and moral" because that is the default response you've provided. The machine has no such ability. If I choose a strapping, muscular fellow rather than a weak, scrawny woman to help me shift a grand piano up the stairs does that reflect my "bias" borne of lower "ideals"? You are suffering from two main problems. 1). You are unable to accept the fact that a machine cannot be taught morality/fairness. 2). You have lost your focus on the issue: Survival. Are fortune cookies handed out on an individual basis? Are the messages within personally chosen? And if they are would they continue being as accurate if the cookie factory workers were to go on strike? ... and you've clearly never worked within survival conditions. The reason is simple and it is at the heart of your biggest problem. It is imperative that you accept the fact that a machine can make no decision WHAT-SO-EVER based upon moral evaluation. I don't know why you are having such difficulties in understanding that. If you ask me the sum of 2 + 2 is "4" the moral response? I don't see anything moral in it but if you do .......
How would you respond to a person who thinks AI is unable to respond to moral evaluations? ChatGPT I understand the concern of those who believe AI is unable to respond to moral evaluations. It's a complex and controversial topic that requires careful consideration. Here are some points to consider when discussing this issue: Lack of Consciousness: AI systems, including language models like myself, do not possess consciousness or self-awareness. As a result, we don't experience emotions or have a sense of morality. We are simply sophisticated algorithms designed to process and generate text based on patterns in data. Programming and Bias: AI systems operate based on the data they are trained on and the algorithms they use. If the training data contains biased information or reflects human prejudices, AI can inadvertently perpetuate those biases in its responses. This can raise ethical concerns and lead to potentially harmful outcomes. Absence of Empathy: AI lacks the ability to empathize with human emotions, which is a fundamental aspect of moral decision-making. While AI can simulate empathy by recognizing emotions and responding appropriately, it doesn't genuinely feel emotions. Objective vs. Subjective Morality: Moral evaluations often involve subjective judgment influenced by cultural, societal, and personal beliefs. AI, as a tool, can assist in decision-making by providing objective information, but determining what is morally right or wrong ultimately remains a human responsibility. Ethical Frameworks: Designers and developers can embed ethical frameworks into AI systems to guide their decision-making processes. This could involve incorporating specific rules, principles, or guidelines aligned with widely accepted moral theories. Limited Understanding: AI, including language models like myself, has limitations in understanding context, nuance, and the complexity of ethical dilemmas. As a result, we may provide responses that appear moral but lack a deep understanding of the underlying issues. Human Responsibility: It's essential to remember that AI is a tool created by humans. The responsibility for ensuring ethical behavior lies with the developers, users, and policymakers who govern its applications. In conclusion, AI's lack of consciousness and empathy can limit its capacity to respond to moral evaluations. While ethical frameworks can be integrated, the ultimate responsibility for ethical decision-making and moral evaluations lies with humans. As AI technology evolves, it is crucial to remain vigilant about potential risks and work towards creating AI systems that align with human values and respect fundamental ethical principles.
Nope, it is just the opposite of what you have been saying. It can respond to moral evaluations and admits to its limitations, needing human oversight. With the simple moral problem that you gave (and now try to complicate), it came up with the correct moral evaluation per my human oversight.
I'm sorry to tell you this but your understanding of the issue is completely buggered. I told you that a machine cannot evaluate morals and the "copy and paste" link you provided proves exactly what I said.
What an excellent counterargument. I gave you what I thought should be the ethical thing to do with your scenario.....my moral evaluation and AI independently says the same thing but it is not a moral evaluation. Got it.
I've stated numerous times that machines cannot be taught morals and your "cut and paste" quote from the source itself proves it. If you cannot accept it then ........
This should be of substantial importance to the topic of this thread for those who are interested in it: https://www.netflix.com/title/81473681 "These nerds are going to kill us all." I watched this last night / early this morning. What moral values do you suppose a young AI might learn if its greatest source of funding is based on winning a military style arms race? *** Well, the nerds deserve a palate cleanser if you choose to watch that, watch this next https://www.netflix.com/title/81473680
I do not fear AI, as long as man controls the on/off switch. But, I suppose it is possible to lose controll such that in order to beat AI, we'd have to shut down the entire world. I"ts kinda like a virus. There needs to be legislation that the creation of AI must come with a master kill switch.
If you've ever worked with ChatGPT-4, you'll know that when it comes to getting summaries, opinions and evaluations on published works/texts, it's wishy washy as hell. I mean, it has no convictions whatsoever, and always seeks to tread the middle ground. THAT is what the programmers programmed in. However, I'm often impressed with it's results. For example, I asked ChatGPT-4 to critique your point: The argument presented touches on some valid concerns about AI, but there are nuances that should be considered. Let's break down and critique the argument: "No matter how sophisticated AI might be, AI is no different from any other programmed computer." Critique: While it's true that AI operates on computers and is, in essence, a program, there's a significant difference between traditional software and advanced AI models. Traditional software operates based on explicit instructions given by programmers, whereas AI, especially machine learning models, learns patterns from data. The sophistication of AI refers to its ability to process vast amounts of data, recognize patterns, and make decisions based on those patterns, often in ways that aren't explicitly programmed. "It can be programmed with bias that reflects the values, attitudes, and politics of the person who creates it." Critique: This is a valid concern. AI models, especially those trained on large datasets, can indeed inherit biases present in the data. However, it's not always a direct result of the "values, attitudes, and politics" of the creator. Sometimes, these biases emerge from historical data or societal biases that the AI has been exposed to. It's also worth noting that there's ongoing research and many initiatives aimed at detecting and mitigating biases in AI. "From that aspect, there is nothing superior about it." Critique: The superiority of AI doesn't stem from its moral or ethical standing but from its computational abilities. AI can process information, recognize patterns, and make decisions at scales and speeds that humans cannot. The argument conflates ethical concerns with computational capabilities. While it's crucial to address and be wary of biases in AI, it doesn't negate the fact that AI has capabilities that can be considered "superior" to traditional computing methods or human processing in specific contexts. In summary, while the argument raises valid concerns about biases in AI, it oversimplifies the nature of AI and its capabilities. It's essential to differentiate between the ethical implications of AI and its computational advantages. Not bad, not bad at all.
As I sit here typing this response, I love the way that AI makes me seem smarter than I really am. It corrects my spelling and offers me choices on grammar. You can ask Google questions and "poof" it starts typing an answer.
As I sit here typing this response, I love the way that AI makes me seem smarter than I really am. It corrects my spelling and offers me choices on grammar. You can ask Google questions and "poof" it starts typing an answer.
