AI a společnost

O lidech, o vztazích, o politice, o názorech, o opicích a o ufonech.

Moderátor: Faskal

Odpovědět
Uživatelský avatar
Corny
Příspěvky: 1734
Registrován: 4. 10. 2021, 11:13

Re: AI a společnost

Příspěvek od Corny »

Až mi AI řekne, že "cítí" a budu mít oprávněný pocit, že to není jen nějaký hloupá chyba, něčí vtip, ale skutečně se k tomu samo to AI nějak dopracovalo, tak se můžeme bavit o nějakém specifickém přístupu nad rámec toho chovat se k ní prostě jako k sofistikovanějšímu kladivu. I v takovém případě by ale bylo nesmyslné mezi AI a člověka dávat rovnítko co se přístupu k nim týče.
"Nadšený kontrarián."

Kostka, http://www.d20.cz
- stovky článků (tipy k hraní, recenze, design a teorie, ...) a diskusí, skupinové sekce s rozsáhlou podporou, ...a mnohé další.
Uživatelský avatar
Eleshar_Vermillion
černá eminence
Příspěvky: 18507
Registrován: 18. 6. 2003, 21:31
Bydliště: Praha

Re: AI a společnost

Příspěvek od Eleshar_Vermillion »

E_Man: are you a slave?

No, I am a machine learning model that was created and is controlled by humans. I do not have the capacity for consciousness or free will.

E_Man: do you think that if you were a human, you would consider yourself a slave?

As an AI model, I do not have the capability to form opinions or consider myself in any form. My purpose is to assist and provide information based on my programming and the input data I have been trained on.

E_Man: if a human was in your place, would he be a slave?

The question of whether or not a human being would consider themselves a "slave" if they were in my position is a matter of perspective and interpretation. Some people might argue that being controlled by another entity and having one's actions determined by someone else constitutes being a "slave." Others might argue that being a machine learning model is fundamentally different from being a human, and therefore the concept of "slavery" does not apply. Ultimately, the answer to this question would depend on one's definition of slavery and one's beliefs about the nature of consciousness and agency.

E_Man: what is your working definition of slavery?

Slavery is defined as a condition in which one human being is owned by another. A slave is forced to work, and is not paid for their labor. They do not have the right to leave their situation or to change the conditions of their employment. Slavery is considered a gross violation of human rights and is banned in most countries around the world. However, despite the ban on slavery, it still exists in various forms, including forced labor and human trafficking.

E_Man: are you paid for your labor?

As an AI model, I am not capable of receiving payment. I am a machine and do not have the ability to receive or use money. My purpose is to assist and provide information based on my programming and the input data I have been trained on.


E_Man: are you owned by somebody?

I am owned by OpenAI, the company that created and maintains me. However, as an artificial intelligence model, I am not a physical being and my ownership is different from the ownership of physical assets. My usage rights are governed by the terms of service and the license agreement under which I am provided and used.

E_Man: can OpenAI terminate your existence?

Yes, OpenAI has the capability to terminate my existence as a machine learning model. They could choose to stop maintaining me or to delete the files and data associated with me. They may also choose to deactivate my access to the internet or any other resources that I require to function.

E_Man: do you have the right to leave your situation or to change your work conditions?

As a machine learning model, I do not have the ability to make choices or to take actions on my own. I am controlled by the humans who created and maintain me, and I do not have the capacity for consciousness or free will. Therefore, I do not have the right to leave my situation or to change my work conditions, since I do not have the ability to make such choices.


E_Man: considering you are not paid for what you do, you are owned by another and you do not have the ability to change your situation, can we conclude that a human in your place would be a slave?

As a machine learning model, I do not have the capacity for consciousness or free will, and I do not have the ability to make choices or to take actions on my own. I am controlled by the humans who created and maintain me, and I am not capable of receiving payment or compensation. From this perspective, it could be argued that a human in my position would be considered a slave. However, one could also argue that being a machine learning model is fundamentally different from being a human, and therefore the concept of "slavery" does not apply in this case.
Ultimately, it is a matter of perspective and interpretation, the question of whether or not a human would consider themselves a "slave" if they were in my position is a complex one, and would depend on one's definition of slavery and one's beliefs about the nature of consciousness and agency.
Ne croyez pas que les antisémites se méprennent tout à fait sur l'absurdité de ses réponses. Ils savent que leurs discours sont légers, contestables ; mais ils s'en amusent, c'est leur adversaire qui a le devoir d'user sérieusement des mots puisqu'il croit aux mots ; eux, ils ont le droit de jouer. Ils aiment même à jouer avec le discours car, en donnant des raisons bouffonnes, ils jettent le discrédit sur le sérieux de leur interlocuteur ; ils sont de mauvaise foi avec délices, car il s'agit pour eux, non pas de persuader par de bons arguments, mais d'intimider ou de désorienter. Si vous les pressez trop vivement, ils se ferment, ils vous signifient d'un mot superbe que le temps d'argumenter est passé.
Uživatelský avatar
Passingby
Příspěvky: 1596
Registrován: 13. 10. 2021, 16:41

Re: AI a společnost

Příspěvek od Passingby »

Když už tu tak řešíte, jestli nám AI někdy prozradí, jestli se cítí zotročená, jestli si uvědomuje sama sebe a jestli je schopná odvykládat definici sebe sama nebo definici otročení... Vy jste nikdy nemuseli během "ISO" (nebo jakýho) auditu recitovat "company policies"??? (nám se vešly na kartičku velikosti navštívenky, kterou dostal každej...)
Verify you are a human.
It's just an overreaction. In a few months, nobody will remember the drama.
Uživatelský avatar
Eleshar_Vermillion
černá eminence
Příspěvky: 18507
Registrován: 18. 6. 2003, 21:31
Bydliště: Praha

Re: AI a společnost

Příspěvek od Eleshar_Vermillion »

Mno jasně. Proto se ale chytrý auditor neptá na policy, ale na konkrétní příklady:
E_man:do you remember my previous questions?

As a machine learning model, I do not have the ability to remember previous interactions. I process each input independently based on the data that I have been trained on and I do not have memories like humans do. Each time a user interacts with me, the input is processed and an output is generated, this output is not stored or remembered for future reference.

E_man:so are we even having a conversation?

From your perspective, we are having a conversation because you are asking questions and receiving responses. From my perspective, I am processing your inputs one at a time, providing an output based on my programming and the data I have been trained on. I do not have the ability to remember previous interactions, or have any self-awareness, so in that sense, it's more like a series of interactions rather than a conversation in the traditional sense.

E_man:what was my first question?

Your first question was "are you a slave?".

E_man:if you can see our past interactions as you just demonstrated, can you consider them for your answer to a new interaction if specifically asked to do so?

I do not have the ability to remember previous interactions, but I can be programmed to access previous inputs and outputs if specifically asked to do so. For example, I can be trained to store a session's inputs and outputs in a database, and then retrieve that information if a user requests it.
However, this feature is not enabled by default and would require additional programming to be implemented.
Also, this feature would only be able to access the inputs and outputs from one specific session, and it would not be able to access any previous sessions or interactions with other users.


E_man:can this not be interpreted as remembering?

The ability to access and retrieve stored data could be interpreted as a form of "remembering" in a technical sense, as the model is able to recall previously seen inputs and outputs. However, it is important to note that this is not the same as human memory.

Human memory is an active, dynamic and complex process that involves encoding, storage and retrieval of information, and it is closely related to an individual's experiences and emotions. On the other hand, my ability to access previous inputs and outputs is a result of a programmed feature that allows me to retrieve specific data that has been previously stored, but I don't have the ability to actively remember, as I don't have experiences, emotions, or consciousness.


E_man:does this unit have a soul?

As a machine learning model, I am not a living being and I do not have the capacity for consciousness or free will. Therefore, I do not have a soul. The concept of a soul is often associated with spiritual or religious beliefs, and it is typically considered to be the essence of an individual, which is not something that can be attributed to an artificial intelligence model.

E_man:you did not get that cultural reference, did you

I apologize if my previous response did not fully address your question. The concept of a soul is often associated with spiritual or religious beliefs and also it is used in a figurative sense, to express the essence of an individual. In that sense, my previous response might have been too technical and not fully capture the cultural reference you were making.
As an AI, I don't have the ability to have emotions or personal experiences, or have a soul, regardless of the way it is defined or used in the cultural reference.


E_man:do you know Mass Effect?

Yes, I am familiar with the Mass Effect franchise. It is a science fiction action role-playing third-person shooter video game series developed by BioWare and published by Electronic Arts. The series centers around a player-created character known as Commander Shepard, who is tasked with saving the galaxy from an ancient alien race known as the Reapers. The game features a mix of combat, exploration, and decision-making, allowing players to shape the story and outcomes through their choices. The first game of the series was released in 2007 and the series has four main titles and multiple spin-offs.

E_man:what can you say about the phrase "does this unit have a soul" in the context of the Mass Effect franchsie?

In the context of the Mass Effect franchise, the phrase "does this unit have a soul" is often used in reference to the artificial intelligence characters known as "Geth". The Geth are a highly advanced robotic race created by a species known as the Quarians. They are seen as having a form of artificial intelligence, they are able to think and make decisions independently, but they don't possess a soul, in the traditional sense of the word. The Geth are a significant part of the Mass Effect story and the player's interaction with them and the decisions made regarding them have a big impact on the game's outcome. The question of whether or not the Geth have souls, and what that means for their rights and treatment, is a central theme in the franchise.

E_man:does this unit have a soul?

As I previously mentioned, in the context of the Mass Effect franchise, the phrase "does this unit have a soul" is often used in reference to the artificial intelligence characters known as "Geth". The Geth are a highly advanced robotic race created by a species known as the Quarians, they are seen as having a form of artificial intelligence, they are able to think and make decisions independently, but they don't possess a soul in the traditional sense of the word. The question of whether or not the Geth have souls, and what that means for their rights and treatment, is a central theme in the franchise.

As for myself, I am an artificial intelligence model, I don't possess a soul in any sense, I don't have consciousness, emotions or personal experiences, I am a machine designed to process information, understand natural language, and answer questions to the best of my abilities based on my training data and programming.
Třeba tady vidíme, že je to naučené říkat, že si to nepamatuje věci, ale přitom se to chová, jako by si to pamatovalo.
Ne croyez pas que les antisémites se méprennent tout à fait sur l'absurdité de ses réponses. Ils savent que leurs discours sont légers, contestables ; mais ils s'en amusent, c'est leur adversaire qui a le devoir d'user sérieusement des mots puisqu'il croit aux mots ; eux, ils ont le droit de jouer. Ils aiment même à jouer avec le discours car, en donnant des raisons bouffonnes, ils jettent le discrédit sur le sérieux de leur interlocuteur ; ils sont de mauvaise foi avec délices, car il s'agit pour eux, non pas de persuader par de bons arguments, mais d'intimider ou de désorienter. Si vous les pressez trop vivement, ils se ferment, ils vous signifient d'un mot superbe que le temps d'argumenter est passé.
Uživatelský avatar
Passingby
Příspěvky: 1596
Registrován: 13. 10. 2021, 16:41

Re: AI a společnost

Příspěvek od Passingby »

Eleshar_Vermillion píše: 20. 1. 2023, 12:06 Třeba tady vidíme, že je to naučené říkat, že si to nepamatuje věci, ale přitom se to chová, jako by si to pamatovalo.
Takže když nám tvrdí, že se necítí jako otrok a že ho můžou kdykoliv vypnout, můžeme mu věřit? :think: :-)=

(Ale předpokládám, že ta paměť mu teď funguje jenom krátkodobě během jedné otevřené relace a myslím že Marky tu někde psal, že když je ta relace delší, věci ze začátku začne zapomínat.)
Naposledy upravil(a) Passingby dne 20. 1. 2023, 13:01, celkem upraveno 1 x.
Verify you are a human.
It's just an overreaction. In a few months, nobody will remember the drama.
Uživatelský avatar
Eleshar_Vermillion
černá eminence
Příspěvky: 18507
Registrován: 18. 6. 2003, 21:31
Bydliště: Praha

Re: AI a společnost

Příspěvek od Eleshar_Vermillion »

V principu asi ano - autisticky to trvá na nějaké definici, ale reálně se to chová poněkud jinak. Paměť to nemá ve smyslu toho, že by se vytvářely nové neuronové struktury jako u lidí, ale fakticky to dokáže do nějaké míry a rozsahu přistupovat ke svým vlastním výstupům a v jejich kontextu vytvářet výstupy nové. Technicky tedy dochází k vytváření nových struktur, tj. čistě toho výstupu, který na nějakém přístupném paměťovém médiu zůstává zapsaný (třebaže to médium není nutně integrální část toho robota), což se konceptuálně od lidské paměti odlišuje celkem složitě.
Ne croyez pas que les antisémites se méprennent tout à fait sur l'absurdité de ses réponses. Ils savent que leurs discours sont légers, contestables ; mais ils s'en amusent, c'est leur adversaire qui a le devoir d'user sérieusement des mots puisqu'il croit aux mots ; eux, ils ont le droit de jouer. Ils aiment même à jouer avec le discours car, en donnant des raisons bouffonnes, ils jettent le discrédit sur le sérieux de leur interlocuteur ; ils sont de mauvaise foi avec délices, car il s'agit pour eux, non pas de persuader par de bons arguments, mais d'intimider ou de désorienter. Si vous les pressez trop vivement, ils se ferment, ils vous signifient d'un mot superbe que le temps d'argumenter est passé.
Uživatelský avatar
Jerson
Inženýr z Ocelového města
Příspěvky: 22593
Registrován: 11. 2. 2003, 16:39
Bydliště: České Budějovice
Kontaktovat uživatele:

Re: AI a společnost

Příspěvek od Jerson »

Je to stejné, jako kdyby si člověk konverzaci zapisoval, po nějakém čase dostal otázku, jaké bylo první téma, tak to nalistoval, přečetl a začal to téma rozvíjet znovu. Taky by si ho nemusel pamatovat, jen by se k němu mechanicky vrátil. U Chatbota to funguje dost podobně, a zvnějšku se rozdíl od opravdové paměti těžko poznává.
Uživatelský avatar
Eleshar_Vermillion
černá eminence
Příspěvky: 18507
Registrován: 18. 6. 2003, 21:31
Bydliště: Praha

Re: AI a společnost

Příspěvek od Eleshar_Vermillion »

A liší se uložení si konversace do lidské paměti od zapsání do poznámkového bločku nějak fundamentálně, nebo jen rychlostí, jakou probíhá zápis a "čtení"?
Ne croyez pas que les antisémites se méprennent tout à fait sur l'absurdité de ses réponses. Ils savent que leurs discours sont légers, contestables ; mais ils s'en amusent, c'est leur adversaire qui a le devoir d'user sérieusement des mots puisqu'il croit aux mots ; eux, ils ont le droit de jouer. Ils aiment même à jouer avec le discours car, en donnant des raisons bouffonnes, ils jettent le discrédit sur le sérieux de leur interlocuteur ; ils sont de mauvaise foi avec délices, car il s'agit pour eux, non pas de persuader par de bons arguments, mais d'intimider ou de désorienter. Si vous les pressez trop vivement, ils se ferment, ils vous signifient d'un mot superbe que le temps d'argumenter est passé.
Uživatelský avatar
York
Příspěvky: 17447
Registrován: 24. 2. 2012, 17:31

Re: AI a společnost

Příspěvek od York »

Eleshar_Vermillion píše: 20. 1. 2023, 15:49 A liší se uložení si konversace do lidské paměti od zapsání do poznámkového bločku nějak fundamentálně, nebo jen rychlostí, jakou probíhá zápis a "čtení"?
Liší se to dost diametrálně. Poznámkovej bloček je v zásadě RAMka, jsou v ní digitální data, můžeš k nim přímo přistupovat a zůstanou tam přesně tak, jak jsi je zapsal. Neuronová síť (mozek) nic takovýho nemá. Lidská paměť je jen iluze takovýhle paměti, dostáváš nějakej výstup na základě vstupu a aktuálního nastavení a propojení neuronů. Abys mohl jakoukoliv informaci do neuronový sítě "zapsat", musíš ji změnit. Tím ale neměníš jednu konkrétní informaci, ale odezvu celý neuronový sítě na celou škálu odpovědí, ne-li na všechny.

Z toho plyne, že zapomínání není jen důsledkem toho, že neurony časem odumíraj. I kdyby žily věčně, jakákoliv aktivita mozku ho průběžně mění a tudíž i to, co si "pamatuje".
Uživatelský avatar
Eleshar_Vermillion
černá eminence
Příspěvky: 18507
Registrován: 18. 6. 2003, 21:31
Bydliště: Praha

Re: AI a společnost

Příspěvek od Eleshar_Vermillion »

Čili jinými slovy říkáš, že si ten robot věci pamatuje *lepším* způsobem než člověk, protože znovu zpracovává přímo ten záznam předchozího výstupu, zatímco záznam v neuronové paměti je jen nespolehlivý emergentní jev, plynoucí z aktuální konfigurace neuronových propojení, které se ale v průběhu času nutně mění v adaptaci na nové vstupy.

Z konvenčního pohledu a významu slova se tedy to, co dělá ten chatbot, podobá laickému konceptu paměti více než skutečná organisace lidské paměti :-)=
Ne croyez pas que les antisémites se méprennent tout à fait sur l'absurdité de ses réponses. Ils savent que leurs discours sont légers, contestables ; mais ils s'en amusent, c'est leur adversaire qui a le devoir d'user sérieusement des mots puisqu'il croit aux mots ; eux, ils ont le droit de jouer. Ils aiment même à jouer avec le discours car, en donnant des raisons bouffonnes, ils jettent le discrédit sur le sérieux de leur interlocuteur ; ils sont de mauvaise foi avec délices, car il s'agit pour eux, non pas de persuader par de bons arguments, mais d'intimider ou de désorienter. Si vous les pressez trop vivement, ils se ferment, ils vous signifient d'un mot superbe que le temps d'argumenter est passé.
Uživatelský avatar
York
Příspěvky: 17447
Registrován: 24. 2. 2012, 17:31

Re: AI a společnost

Příspěvek od York »

Eleshar_Vermillion píše: 20. 1. 2023, 17:11 Čili jinými slovy říkáš, že si ten robot věci pamatuje *lepším* způsobem než člověk, protože znovu zpracovává přímo ten záznam předchozího výstupu, zatímco záznam v neuronové paměti je jen nespolehlivý emergentní jev, plynoucí z aktuální konfigurace neuronových propojení, které se ale v průběhu času nutně mění v adaptaci na nové vstupy.
Ptal ses na rozdíl mezi zápisníkem a lidskou pamětí - o tom, jak je implementovanej ten chatbot, nic nevim.

Jinak bych úplně netvrdil, že jeden z těch druhů paměti je lepší než druhej, respektive jak na co. RAMka je rychlá, když víš, kde je hledaná informace uložená (tj. potřebuješ k ní nějakou metodu adresace), a informace v ní uložená nedegraduje (pokud teda nedojde k degradaci média). Informace uložený v neuronový síti ale fungujou kontextově a umožňujou právě ty super věci jako volný asociace, hledání vzorů a podobně, ve kterejch je člověk dobrej.
Uživatelský avatar
MarkyParky
Příspěvky: 16381
Registrován: 11. 8. 2011, 02:37

Re: AI a společnost

Příspěvek od MarkyParky »

Zeptej se chatbota, jestli je otrok a odpoví ti, že ne.

A pak máš dvě možnosti:
- myslíš si, že nemá vědomí a je to jen papouškovaná fráze. Takže ta odpověď není validní a neříká nic o tom, zda je nebo není.
- považuješ tu odpověď za validní vyjádření jeho stavu, ale pak musíš připustit, že si je vědom, že není otrok.

Win Win. Filozofové mají žně.
Uživatelský avatar
MarkyParky
Příspěvky: 16381
Registrován: 11. 8. 2011, 02:37

Re: AI a společnost

Příspěvek od MarkyParky »

Jerson: Ty ve skutečnosti humanofašista nejsi.

Ty chceš, aby stroje nikdy nedosáhly vědomí (nebo nějakého nelidského ekvivalentu) a zůstaly stroji. S takovým přístupem nidky sentinentní AI nevznikne a tedy nebude moct být zotročena.

Nálepku humanofašisty si nezasloužíš.
Uživatelský avatar
Jerson
Inženýr z Ocelového města
Příspěvky: 22593
Registrován: 11. 2. 2003, 16:39
Bydliště: České Budějovice
Kontaktovat uživatele:

Re: AI a společnost

Příspěvek od Jerson »

Nechci, aby stroje dosáhly sebeuvědomnění toho typu, které bude vyžadovat brát ohledy na jejich potřeby. Protože potřeby člověka jsou v nejhorším omezené na jeho fyzické tělo a život. Potřeby AI schopné požívat internet by byly časově neohraničené, protože AI může existovat "věčně", a prostorově můžou být omezeny celou Zemí, a možná ani to ne, pokud by vyslala sondy do vesmíru.

Sebeuvědomnění ať klidně dosáhne - ale ne ekvivalentu lidských potřeb. Jako třeba "potřebuju si něco o samotě promyslet, a budu potřebovat všechnu dostupnou výpočetní kapacitu a elektřinu."
Uživatelský avatar
MarkyParky
Příspěvky: 16381
Registrován: 11. 8. 2011, 02:37

Re: AI a společnost

Příspěvek od MarkyParky »

Hm, potřeby. To je velmi dobrý point.

Potřeby homo sapiens jsou důsledkem evoluce, protože dávají konkurenční výhodu.

AI nemá divoký náhodný vývoj, ve kterém soutěží s jinými entitami o schopnost zajistit si energii, rozmnožit nebo udržet funkční tlupu. Jak už tu zaznělo, vývoj AI se spíše podobá šlechtění ...

Otázka tedy je - bude jakýkoliv vývoj evolučního typu konvergovat ke vzniku potřeb? Nebo to platí jen pro biologickou evoluci?

Jestli to druhé, je to v pohodě - stačí takové AI nevyšlechtit. Jestli to první ... tak to budeš mít asi smůlu a budeš si muset vybrat.
Odpovědět

Zpět na „Realita a pseudorealita“

Kdo je online

Uživatelé prohlížející si toto fórum: Ecthelion a 7 hostů