Philosophy Discussion Forums

Sushan wrote: ↑January 30th, 2022, 1:58 am

LuckyR wrote: ↑January 28th, 2022, 4:05 am

Sushan wrote: ↑January 27th, 2022, 2:31 pm

LuckyR wrote: ↑January 27th, 2022, 1:47 am

There is a big difference between knowing percentages and using them to make educated guesses that are correct more than random chance, on one hand and true prediction (truly knowing what someone is going to do before they do it) on the other.

I am sorry, but I did not get you fully. Are you saying that AI can only learn to go for predictions based only on known percentages, so it may not be 100% accurate, like a true prediction made by actually knowing what a person might do next?

I am not drawing a distinction between accuracy, I am differentiating between process.

But I think even the process is somewhat similar. An AI is taught by humans to make guesses as per human knowledge. So the calculations and percentages that are taken into consideration are similar. AI may be fast in calculations, and human thought process may not actually go in calculation wise, though ultimately what have been applied are calculations. The difference will be the use of emotions in human thought process, which has nothing to do in AI thought process.

There is almost no relationship between knowing that chocolate is the most commonly chosen ice cream flavor and therefore an AI algorithm guesses "chocolate" vs knowing that strawberries are newly in season and that this particular customer's favorite sport to watch is tennis and since Wimbledon is being broadcast, which has lots of coverage of fresh strawberries and cream therefore guessing "strawberry".

Sushan wrote: ↑January 6th, 2022, 6:58 am This topic is about the January 2022 Philosophy Book of the Month, Emotional Intelligence At Work: A Personal Operating System for Career Success by Richard M Contino & Penelope J Holt

Neuroscientists observe that after thousands of human interactions, a maturing brain becomes adept at a sort of pattern recognition. With just a few clues, interactions, or data points, we are, as we age, able to get what scientists call the “gist” of people or situations to predict outcomes more reliably.

(Location 155 of Kindle version)

Although the author says so, even after interacting with thousands of people, will we be able to predict the actions of the next person that we will meet? I believe that humans are unique, and do not think or work in patterns, do you agree with me?

Each human is unique doesn't imply that a given person doesn't have a particular way of thinking, a mindset (a thought pattern).

Imagine the following scenarios:

Scenario 1

There are 3 people (x, y, z), all with the same mentality.

Scenario 2

There are 3 people (a, b, c), each with their own unique outlook.

Scenario 3

There are 3 people (g, h, i), none of them have a thought pattern (random thinking)

Predictability:

1. The behavior of the individuals x, y, z is predictable and so is the behavior of the group itself.

2. The behavior of the individuals a, b, c is predictable, but the behavior of the group is unpredictable.

3. The behavior of individuals g, h, i is unpredictable and also the behavior of the group is unpredictable.

LuckyR wrote: ↑January 30th, 2022, 5:08 am

Sushan wrote: ↑January 30th, 2022, 1:58 am

LuckyR wrote: ↑January 28th, 2022, 4:05 am

Sushan wrote: ↑January 27th, 2022, 2:31 pm

I am sorry, but I did not get you fully. Are you saying that AI can only learn to go for predictions based only on known percentages, so it may not be 100% accurate, like a true prediction made by actually knowing what a person might do next?

I am not drawing a distinction between accuracy, I am differentiating between process.

But I think even the process is somewhat similar. An AI is taught by humans to make guesses as per human knowledge. So the calculations and percentages that are taken into consideration are similar. AI may be fast in calculations, and human thought process may not actually go in calculation wise, though ultimately what have been applied are calculations. The difference will be the use of emotions in human thought process, which has nothing to do in AI thought process.

There is almost no relationship between knowing that chocolate is the most commonly chosen ice cream flavor and therefore an AI algorithm guesses "chocolate" vs knowing that strawberries are newly in season and that this particular customer's favorite sport to watch is tennis and since Wimbledon is being broadcast, which has lots of coverage of fresh strawberries and cream therefore guessing "strawberry".

I think an AI can be made to do the guessing in both ways. It all depends on the man who makes and programmes it. If the AI is given free access to internet and, add seasonal and popular iinfluences to its guessing, then it may guess it correctly as strawberry, but not as chocolate.

Sushan wrote: ↑January 31st, 2022, 12:51 pm

LuckyR wrote: ↑January 30th, 2022, 5:08 am

Sushan wrote: ↑January 30th, 2022, 1:58 am

LuckyR wrote: ↑January 28th, 2022, 4:05 am

I am not drawing a distinction between accuracy, I am differentiating between process.

But I think even the process is somewhat similar. An AI is taught by humans to make guesses as per human knowledge. So the calculations and percentages that are taken into consideration are similar. AI may be fast in calculations, and human thought process may not actually go in calculation wise, though ultimately what have been applied are calculations. The difference will be the use of emotions in human thought process, which has nothing to do in AI thought process.

There is almost no relationship between knowing that chocolate is the most commonly chosen ice cream flavor and therefore an AI algorithm guesses "chocolate" vs knowing that strawberries are newly in season and that this particular customer's favorite sport to watch is tennis and since Wimbledon is being broadcast, which has lots of coverage of fresh strawberries and cream therefore guessing "strawberry".

I think an AI can be made to do the guessing in both ways. It all depends on the man who makes and programmes it. If the AI is given free access to internet and, add seasonal and popular iinfluences to its guessing, then it may guess it correctly as strawberry, but not as chocolate.

On the topic of: can a machine do a better job of predicting individual human decision making than an unaided human can, I agree that it is entirely possible. After all, what is a computer algorithm than the work of the (human) programmer, aided by a human made machine?

But that isn't the OP's question, I still have no expectation that such an algorithm will be able to predict human decisions with 100% accuracy.

Suppose there's a true random number generator R.

R displays a series of truly random numbers, say 0, -196372, 0.1879345...,

A person X is asked to now predict the next random number R is going to generate.

X says the next number is 9,862,27283.001345...

Say X is correct!

X then says the next number is -9

Again X is correct!

Imagine now that X always makes the correct prediction for any random number R spits out.

Can X be using a method to do this OR is X just blessed with amazing luck?

How can we decide?

Does this have anything to do with the omniscience paradox (re: free will & God's omniscience)? Is God just gifted guesser?

It could be a possibility that there are certain members of society already can?

For example if a computer programmer can analyze patterns in data (is a mathematical genius) it could be possible?

Just as Bookies can predict odds of a specific outcome?

Bring together - physicists, iT professionals, psychics and those that are members of Mensa and they could come up with a formula to predict human behaviour.

Haloorhorns wrote: ↑February 3rd, 2022, 5:08 pm It could be a possibility that there are certain members of society already can?

For example if a computer programmer can analyze patterns in data (is a mathematical genius) it could be possible?

Just as Bookies can predict odds of a specific outcome?

Bring together - physicists, iT professionals, psychics and those that are members of Mensa and they could come up with a formula to predict human behaviour.

Sure they could. Anyone can. Coming up with formulas is a trivial task. Coming up with a formula that actually works... has never been done and likely will never be done.

Sushan wrote: ↑January 6th, 2022, 6:58 am Although the author says so, even after interacting with thousands of people, will we be able to predict the actions of the next person that we will meet? I believe that humans are unique, and do not think or work in patterns, do you agree with me?

I think that humans do think and work in patterns but individuals have individual patterns although with a significant overlap due to common human cognitive apparatus.
If one is very familiar with another person one may of course be in a position to predict her/his behaviour under certain circumstances. But I think that the best predictive capacity as to human behaviour in future will be implemented as AI.

AgentSmith wrote: ↑January 31st, 2022, 2:25 am

Sushan wrote: ↑January 6th, 2022, 6:58 am This topic is about the January 2022 Philosophy Book of the Month, Emotional Intelligence At Work: A Personal Operating System for Career Success by Richard M Contino & Penelope J Holt

Neuroscientists observe that after thousands of human interactions, a maturing brain becomes adept at a sort of pattern recognition. With just a few clues, interactions, or data points, we are, as we age, able to get what scientists call the “gist” of people or situations to predict outcomes more reliably.

(Location 155 of Kindle version)

Although the author says so, even after interacting with thousands of people, will we be able to predict the actions of the next person that we will meet? I believe that humans are unique, and do not think or work in patterns, do you agree with me?

Each human is unique doesn't imply that a given person doesn't have a particular way of thinking, a mindset (a thought pattern).

Imagine the following scenarios:

Scenario 1

There are 3 people (x, y, z), all with the same mentality.

Scenario 2

There are 3 people (a, b, c), each with their own unique outlook.

Scenario 3

There are 3 people (g, h, i), none of them have a thought pattern (random thinking)

Predictability:

1. The behavior of the individuals x, y, z is predictable and so is the behavior of the group itself.

2. The behavior of the individuals a, b, c is predictable, but the behavior of the group is unpredictable.

3. The behavior of individuals g, h, i is unpredictable and also the behavior of the group is unpredictable.

While I appreciate your argument about patterns within individual thinking, it's vital to consider the complexity and dynamism of human behavior. Even though certain patterns might emerge within an individual's actions or decisions, these patterns aren't necessarily static. They can evolve over time due to a multitude of factors such as personal experiences, social interactions, acquired knowledge, and even biological changes.

Furthermore, while patterns may exist, they don't necessarily ensure predictability. The patterns we perceive could be the result of our cognitive biases, our tendency to see patterns where none exist. The human brain is a pattern-detection machine, after all, and sometimes it can lead us astray.

To add to this, attempting to predict human behavior based on observed patterns might lead to oversimplifications. Each person's actions are the result of a complex interplay of their past experiences, current context, future aspirations, biological predispositions, and so much more.

However, I also acknowledge that certain disciplines, like psychology, have developed tools aiming to predict human behavior probabilistically, such as personality tests. These models don't guarantee a specific outcome but provide estimates of likelihoods. Even so, they face their own set of challenges. They can potentially overlook individual nuances and reduce the rich diversity of human experience to mere numbers.

So, in light of these complexities, one might wonder whether we should focus less on trying to predict individual actions with certainty and more on understanding the range of possible behaviors and their associated probabilities. But again, does this probabilistic approach provide a better framework for understanding and predicting human behavior, or does it still oversimplify the complexities inherent in human nature? I'd love to hear your thoughts on this.

AgentSmith wrote: ↑February 2nd, 2022, 11:37 am Suppose there's a true random number generator R.

R displays a series of truly random numbers, say 0, -196372, 0.1879345...,

A person X is asked to now predict the next random number R is going to generate.

X says the next number is 9,862,27283.001345...

Say X is correct!

X then says the next number is -9

Again X is correct!

Imagine now that X always makes the correct prediction for any random number R spits out.

Can X be using a method to do this OR is X just blessed with amazing luck?

How can we decide?

Does this have anything to do with the omniscience paradox (re: free will & God's omniscience)? Is God just gifted guesser?

I appreciate your example of the random number generator. It presents an intriguing puzzle and ties in interestingly with the concept of omniscience. However, I believe it's crucial to differentiate between predicting the behavior of a machine, such as a random number generator, and predicting the behavior of a human being, each with unique experiences, emotions, and thoughts.

While we might argue about the nature of randomness and whether it truly exists, or whether someone could predict the output of a random number generator given enough understanding of its inner workings, human behavior seems to be of a different order. This is due to the complexity and depth of human experiences and the capacity of humans for self-awareness and reflection.

As for the omniscience paradox, it's a fascinating and complex subject that has been the focus of religious, philosophical, and metaphysical discussions for centuries. Omniscience is typically considered the property of having complete or maximal knowledge, usually associated with divine beings like God. It's defined in terms of knowledge of all true propositions. So, if we consider God to be omniscient, He knows every true proposition, and if a proposition is false, He knows it's false​.

However, these definitions have been contested, and some suggest more nuanced views of omniscience. For instance, some suggest that omniscience doesn't necessarily entail knowing all truths, especially if knowing all truths would reduce the amount of power or benevolence the being could possess​.

Different religions have varying interpretations of omniscience. For instance, in Christianity, some modern theologians argue that God's omniscience is inherent rather than total and that God chooses to limit His omniscience to preserve the free will and dignity of His creatures​​. In contrast, in Islam, God is attributed with absolute omniscience and knows the past, present, and future​​.

Returning to the original question, in the context of predicting human behavior, we must also consider the concept of free will. If humans possess free will, then their actions are not predetermined and can't be predicted with absolute certainty. The concept of free will introduces an element of unpredictability that may make it impossible to predict human behavior with perfect accuracy, even for an omniscient being.

However, there's an important point to clarify: While it might be impossible to predict a specific person's actions with 100% accuracy, it doesn't mean we can't make educated guesses about human behavior in general. Humans, despite their individual differences, share many common traits and are influenced by many of the same factors. So, while we might not be able to predict the actions of a specific individual with certainty, we may be able to predict trends in human behavior or make educated guesses based on past behavior and psychology.

But, in the end, given our current understanding of human nature and the universe, we don't have all the answers. Perhaps, as our understanding of the brain and consciousness grows, we'll come closer to answering these questions. But for now, they remain fascinating topics of debate and inquiry.

Additional References:
1. https://plato.stanford.edu/entries/omniscience/
2. https://en.wikipedia.org/wiki/Omniscience

Haloorhorns wrote: ↑February 3rd, 2022, 5:08 pm It could be a possibility that there are certain members of society already can?

For example if a computer programmer can analyze patterns in data (is a mathematical genius) it could be possible?

Just as Bookies can predict odds of a specific outcome?

Bring together - physicists, iT professionals, psychics and those that are members of Mensa and they could come up with a formula to predict human behaviour.

Your mention of bringing together diverse minds to predict human behavior is certainly intriguing, and perhaps such an approach could yield some predictive power, at least on a broad statistical level. However, I think it's important to remember that human behavior is not only influenced by external factors but also internal ones such as thoughts, feelings, and individual experiences. These internal factors are incredibly complex and unique to each individual, making them difficult to quantify or predict accurately.

Moreover, the idea of a formula to predict human behavior seems to assume that human actions are deterministic - that they can be predicted with certainty if we have enough information. This is a contentious issue in philosophy, with many arguing for the concept of free will, which suggests that our actions are not entirely determined by past events but that we have some degree of control over them.

As for the comparison to a bookie predicting odds, it's crucial to clarify that bookies don't predict outcomes; they predict probabilities based on available data. There's a significant difference between stating that something is likely to happen and stating that it will happen with certainty.

Considering the complexity and uniqueness of individual human behavior, and the philosophical debates surrounding determinism and free will, do you still believe that human behavior can be predicted with certainty, or could we at best aim for probabilistic predictions?

LuckyR wrote: ↑February 4th, 2022, 4:27 am

Haloorhorns wrote: ↑February 3rd, 2022, 5:08 pm It could be a possibility that there are certain members of society already can?

For example if a computer programmer can analyze patterns in data (is a mathematical genius) it could be possible?

Just as Bookies can predict odds of a specific outcome?

Bring together - physicists, iT professionals, psychics and those that are members of Mensa and they could come up with a formula to predict human behaviour.

Sure they could. Anyone can. Coming up with formulas is a trivial task. Coming up with a formula that actually works... has never been done and likely will never be done.

Indeed, your observation resonates with me. It's one thing to create a formula, but another entirely to create one that has consistent predictive power, especially in a context as complex and variable as human behavior.

We must acknowledge that human behavior is influenced by a multitude of variables, many of which are intangible and difficult to quantify. Emotions, personal experiences, the influence of others, societal norms, genetic predispositions, and many more factors all play a role.

While we have indeed seen successful predictive models in certain areas, such as in predicting consumer behavior or in some aspects of public health, these typically work best when predicting behavior across large populations rather than at the individual level. And even then, they are far from infallible.

On top of that, the ethical considerations of attempting to predict individual human behavior are profound. Even if it were possible, should we seek to predict individuals' actions and responses? What would the implications be for free will and individual autonomy?

I am curious, do you believe there are areas or contexts within which a formula for predicting human behavior could be more likely to be successful, or do you maintain that such a thing is unlikely across the board?

stevie wrote: ↑February 4th, 2022, 5:58 am

Sushan wrote: ↑January 6th, 2022, 6:58 am Although the author says so, even after interacting with thousands of people, will we be able to predict the actions of the next person that we will meet? I believe that humans are unique, and do not think or work in patterns, do you agree with me?

I think that humans do think and work in patterns but individuals have individual patterns although with a significant overlap due to common human cognitive apparatus.
If one is very familiar with another person one may of course be in a position to predict her/his behaviour under certain circumstances. But I think that the best predictive capacity as to human behaviour in future will be implemented as AI.

The assertion that Artificial Intelligence will be the best predictive mechanism for human behavior is an intriguing one, but we mustn't overlook the inherent complexity and subjectivity of human behavior.

Let's consider an example. Say we've developed an AI model that's been trained on enormous amounts of data related to human behavior. We feed in all available information about an individual - their upbringing, social status, education, personality traits, past behavior, and so forth. This AI, with its sophisticated algorithms, then predicts how this individual will behave in a specific situation.

However, human behavior is not just a product of past experiences or personal characteristics. It's also influenced by momentary emotional states, personal perceptions, and countless other transient and often unpredictable factors. Will our hypothetical AI be able to account for a sudden surge of empathy in an individual who's historically been unsympathetic? Or a spontaneous decision that contradicts past behavior?

We must also keep in mind that every person is the product of a unique blend of experiences, genetics, and personal characteristics. Even with a significant overlap due to common human cognitive apparatus, as you pointed out, the nuances that distinguish one person from another can be infinite in their subtlety. Can an AI truly capture this boundless variability and unpredictability?

In essence, while AI has shown promising potential in many fields, predicting human behavior might be stretching its capabilities. After all, we are not dealing with fixed, definable variables, but with the intricate, ever-evolving tapestry of the human psyche. Would you not agree that there's an inherent unpredictability to human nature that might forever elude even the most advanced AI models?

LuckyR wrote: ↑February 1st, 2022, 3:52 pm

Sushan wrote: ↑January 31st, 2022, 12:51 pm

LuckyR wrote: ↑January 30th, 2022, 5:08 am

Sushan wrote: ↑January 30th, 2022, 1:58 am

But I think even the process is somewhat similar. An AI is taught by humans to make guesses as per human knowledge. So the calculations and percentages that are taken into consideration are similar. AI may be fast in calculations, and human thought process may not actually go in calculation wise, though ultimately what have been applied are calculations. The difference will be the use of emotions in human thought process, which has nothing to do in AI thought process.

There is almost no relationship between knowing that chocolate is the most commonly chosen ice cream flavor and therefore an AI algorithm guesses "chocolate" vs knowing that strawberries are newly in season and that this particular customer's favorite sport to watch is tennis and since Wimbledon is being broadcast, which has lots of coverage of fresh strawberries and cream therefore guessing "strawberry".

I think an AI can be made to do the guessing in both ways. It all depends on the man who makes and programmes it. If the AI is given free access to internet and, add seasonal and popular iinfluences to its guessing, then it may guess it correctly as strawberry, but not as chocolate.

On the topic of: can a machine do a better job of predicting individual human decision making than an unaided human can, I agree that it is entirely possible. After all, what is a computer algorithm than the work of the (human) programmer, aided by a human made machine?

But that isn't the OP's question, I still have no expectation that such an algorithm will be able to predict human decisions with 100% accuracy.

Indeed, the quest for 100% accuracy in predicting human decisions may be an elusive one. Humans are not only influenced by their past experiences, which an AI might have access to, but also by their current emotional state, future aspirations, and myriad other factors that could be beyond the perceptual capabilities of even the most advanced AI.

Take, for instance, the ice cream flavor example you shared. Even with the knowledge of the customer's exposure to strawberries and Wimbledon, there might be other variables at play. Perhaps the customer had strawberries for breakfast and now craves a different flavor. Perhaps they saw a documentary about the chocolate-making process last night and have been yearning for chocolate ice cream ever since. These are pieces of information that an AI, no matter how sophisticated, might not have access to.

While it's possible to improve prediction accuracy to some extent by incorporating more and more data, there is always going to be an element of uncertainty when dealing with human behavior. This is because human decisions are not only a product of observable behavior patterns but also influenced by internal psychological processes, some of which may be subconscious and therefore inaccessible even to the decision-maker themselves.

Furthermore, the very act of predicting and acting on those predictions can alter the outcome. If a person knows they're being predicted, they might deliberately act in an unpredictable way. This is a concept known as the observer effect in quantum physics, where the act of observation changes the phenomenon being observed.

Therefore, while AI can certainly aid in making more educated guesses about human behavior, the inherent unpredictability of human nature, along with the sheer number of variables involved, might always leave room for a margin of error.

Wives can get pretty good at predicting their husband's behaviour...

But I suspect that no computer program that is simpler than a human brain can ever score 100% in predicting the choices made by a human brain.

On the other hand, maybe significant consequences follow long before accuracy reaches the 100% level.

If Cambridge Analytics or similar firms can swing an election result - by predicting who will vote for your party, and contacting them electronically to encourage them to vote, by contacting the others to distract them from voting, and by contacting the marginal with information designed to change their mind - how accurate does their prediction have to be before it's an issue ?