[Previous] Discussion about Animal Rights and Popper | Home | [Next] The Cambridge Declaration on Consciousness

Animal Rights Issues Regarding Software and AGI

Claim: Animal rights may be refuted by advanced Critical Rationalist (CR) epistemology, including the jump to universality, but most people (pro or anti animal rights) haven’t read and understood The Beginning of Infinity and have a different view of epistemology. Given that ignorance of CR, their belief in animal rights is reasonable. And their failure to understand my questions and challenges of their beliefs is also reasonable. (This claim is based on a comment by TheRat on Discord.)

I disagree with that claim. The purpose of this post is to restate my main question/challenge for animal rights and then to argue that it should be understandable, and be seen as an issue worth answering, by someone who has never heard of CR. The issue is related to software not CR. I will further claim that a non-programmer should be able to understand the question/problem/issue and see that it matters (even though he’ll have a hard time reaching a conclusion about the answer without being able to understand code).

Note: I do have other arguments against animal rights which rely on CR.

The Programmer’s Challenge to Animal Rights

Claim: Animals are complex robots. Humans are different because they have general intelligence – the thing that AGI (Artificial General Intelligence) researchers are trying to program but haven’t yet been able to. All known and documented animal behavior is compatible with animals lacking general intelligence (example).

Animals are built with with different materials (more carbon, less metal). This difference is irrelevant. Similarly, the “artificial” in Artificial General Intelligence doesn’t matter either.

Animals are fundamentally similar to a self-driving car, to board game playing software in a robot with (or without) an arm that can move the pieces around the board, and to “AI” controlled video game characters. Those, like all human-written software that exists today, are all examples of non-AGI (non-general intelligence) algorithms. And the lack of a physical body in some cases important (a robot body could be built and added without changing the intelligence of the software).

Brains of both animals and humans are universal classical computers (Turing complete), just like Macs and iPhones, which run software. The relevant differences are software algorithm differences. People who deny this are ignorant and/or unscientific.

Further Explanation

All software we know how to write today is inadequate to achieve general intelligence. So to claim animals have moral rights like humans, people should argue that animals do things which fundamentally differ from current software. So far I have been unable to find any serious attempt to do this.

Alternatively, someone could come up with a distinguishing feature of software algorithms other than having or lacking general intelligence, show that some animals have that feature, and explain why that feature has moral relevance. I’ve also been unable to find any serious attempt to do this.

Whether general intelligence has moral relevance is non-obvious. Regardless, a reasonable person should agree it might have major moral relevance and therefore this is an issue worth investigating for those curious about animal rights. If there is no animal rights literature trying to do this sort of analysis, and addressing these issues, that’s a significant gap in their arguments.

People denying that general intelligence has moral relevance should specify what else humans have, which robots lack, which they think has moral relevance. A common answer to that is the capacity to suffer. I have been unable to find any animal rights literature that tries to differentiate humans or AGIs from self-driving cars and non-AGI software in terms of ability to suffer. What is it about a human’s software, what trait matters other than general intelligence, that grants the capacity to suffer? If they answered that, then we could investigate whether animal software has that trait or not.

I think capacity to suffer is related to general intelligence because suffering involves making value judgments like not wanting a particular outcome or thinking something is bad. Suffering involves having preferences/wants which you then don’t get. I don’t think it’s possible without the ability to consider alternatives and make value judgments about which you prefer, which requires creative thought and the ability to create new knowledge, think of new things. This is a very brief argument which I’m not going to elaborate on here. My main goal is to challenge animal rights advocates. What is their position on this matter and where are their arguments?

What I’ve mostly found is that people don’t want to think about computer algorithms. They don’t know how to program and they aren’t scientists. They don’t know (or deny without educated arguments) that brains are literally universal classical computers (Turing complete), that information and computation is part of physical reality and physics, that human minds are literally equivalent to some sort of software, and other things like that. That’s OK. Not everyone is an expert.

That’s why I’ve been asking (see the comments in addition to the post) to be referred to literature from someone who does know how to program, understands some of these basic issues, and then makes a case for animal rights. Where are the people with relevant expertise about computers and AGI who favor animal rights and write arguments? I can’t find any. That’s bad for the case for animal rights!

Note: My relevant views on AGI are mainstream for the field. I disagree with the mainstream views in the AGI field on some advanced details, but the basic stuff I’m discussing here is widely agreed on. That doesn’t prove it’s true or anything, but a mainstream view merits some analysis and argument rather than being ignored. (Even obscure views often merit a reply, but I won’t get into that.) If animal rights advocates have failed to consider mainstream AGI ideas, that’s bad.

Consciousness

Besides suffering and general intelligence, the other main trait brought up in animal rights discussions is consciousness. If animals are conscious, that gives them moral value. These three traits are related, e.g. consciousness seems to be a prerequisite of suffering, and consciousness may be a prerequisite or consequence of general intelligence.

What computations, what information processing, what inputs or outputs to what algorithms, what physical states of computer systems like brains indicates or is consciousness? I have the same question for suffering too.

Similar questions can be asked about general intelligence. My answer to that is we don’t entirely know. We haven’t yet written an AGI. So what should we think in the meantime? We can look at whether all animal behavior is consistent with non-AGI, non-conscious, non-suffering robots with the same sorts of features and design as present day software and robots that we have created and do understand. Is there any evidence to differentiate an animal from non-AGI software? I’m not aware of any, although I’ve had many people point me to examples of animal behavior that are blatantly compatible with non-AGI programming algorithms. Humans are different because lots of their behavior is not explainable in terms of current software algorithms. Humans create new knowledge, e.g. about spaceships and vaccines, that isn’t programmed in their genes. And humans do that regarding many different topics, seemingly all, hence the idea of “general” intelligence. I have yet to see evidence that any animal does that on even one topic, let alone generally.

Many of the arguments about consciousness involve the rejection of what I regard as science. E.g. they advocate dualism – they claim that there is something other than the material world. They claim that consciousness is a fundamental, non-physical part of reality. They deny that physics can explain and account for everything that exists.

I regard dualism as bad philosophy but I won’t go into that. I’ll just say that if the case for animal rights relies on the rejection of modern physics and the scientific-materialist view of the world, they’ve got a serious problem which they should address. Where can I read literature telling me why I should change my view of science and accept claims like theirs, which addresses the kind of doubts an atheist who believes in objective physical reality would have? I haven’t gotten any answers to that so far. Instead I’m told assertions which I regard as factually false, e.g. that information is not physical. People who say things like that seem to be unfamiliar with standard views in physics (example paper).

The Argument for Conservatism

Animal rights advocates claim that, if in doubt, we should err on the side of caution. If the science and philosophy of mind isn’t fully figured out, then we should assume animals have moral value just in case they do. Even if there’s only a 1% chance that animals have rights, it’s a bad idea to slaughter them by the millions. I agree.

Pro-life (anti-abortion) advocates make the same argument regarding human fetuses. The science and philosophy aren’t fully settled, so when in doubt we should avoid the chance of murdering millions of human beings, even if it’s a low chance. I agree with that too. I think most animal rights advocates disagree with that or refuse to take it into account so that they can favor abortion. I think this indicates some political bias and double standards. I imagine there are some pro-life animal rights activists, but I think most aren’t, which I think is screwy.

Despite agreeing with these arguments, I’m pro-abortion and pro-slaughtering-farm-animals. The reason I favor abortion is I don’t have any significant doubt about whether a 3 month old fetus, which doesn’t not yet have a brain with electrical activity, is intelligence. I haven’t carefully researched the scientific details about abortion (I would if I was actually deciding the law), but from what I’ve seen, banning third trimester abortions is a reasonable and conservative option.

The reason I favor slaughtering cows is that I have no significant doubt about whether a cow has general intelligence. I’ve seen zero indicators that it does, and I’ve debated many people about this, asked many animal rights advocates for things to read which argue their case, asked for examples of animals doing things which are different than what a non-AGI robot could do, and so on. The total lack of relevant counter-argument from the other side is just the same as with abortion and is about equally conclusive. When all the arguments go one way, one can reasonably reach a conclusion and act on it instead of endlessly doubting. (When argument X has logical priority over Y, then Y is excluded from “all the arguments”. And when argument P is conclusively refuted by argument Q, then P is excluded from “all the arguments”.)

My Expertise

Because I’m asking for arguments from someone familiar with software and AGI rather than from just anyone, I think it’s fair that I share my own background.

I’m a philosopher and programmer. My speciality is epistemology (the philosophy of knowledge, including how to think, learn and reason, and how to evaluate ideas and arguments). I study and contribute to the Critical Rationalist epistemology of Karl Popper and David Deutsch, which I believe is important to making progress on AGI. David Deutsch, a physicist, philosopher and programmer, was my mentor and taught me a lot about philosophy and physics. He’s an award-winning pioneer of quantum computing, a Fellow of the Royal Society, and an author.

I’m a professional programmer with over a decade of work experience, but the software I work on isn’t related to AGI. I’ve read books about AI, watched talks, learned and coded some of the algorithms, talked with people in the field, etc.

Conclusion

Non-programmer animal rights advocates ought to be able to see that someone, some expert, should address the issue of whether humans are animals are differentiated by general intelligence. They should argue that animals have general intelligence (or argue that humans don’t have it) or explain some other sort of software/algorithm/code difference between animals and present day, non-AGI robots and software. If no one can do that and address the computational issues, the remaining option in favor of animal rights is to reject science.

I’m seeking thoughtful, competent written arguments addressing these issues. Blog posts are OK, not just academic material. I challenge anyone who favors animal rights to refer me to such literature in the comments below.


Elliot Temple on November 20, 2019

Messages (10)

https://en.wikipedia.org/wiki/Animal_consciousness#Grey_Parrots

> In 2011, research led by Dalila Bovet of Paris West University Nanterre La Défense, demonstrated grey parrots were able to coordinate and collaborate with each other to an extent. They were able to solve problems such as two birds having to pull strings at the same time to obtain food. In another example, one bird stood on a perch to release a food-laden tray, while the other pulled the tray out from the test apparatus. Both would then feed. The birds were observed waiting for their partners to perform the necessary actions so their behaviour could be synchronized. The parrots appeared to express individual preferences as to which of the other test birds they would work with.[154]

This is a good example of the kind of thing written by people who don't seem to know how to code or have a basic conception of what non-AGI software can easily do (such as everything in that paragraph and far more).


curi at 4:02 PM on November 20, 2019 | #14411 | reply | quote

One of the issues people don't get is the *hardware independence* of computation. They look at hardware features of brains and laptops, see differences, and conclude stuff is different. But software is separate from hardware, this is well known, they have no counter-arguments, they're just ignorant. This is the kind of thing that makes talking to people about this hard when they lack background knowledge re computers.

People also don't get that if all brains are computers, including human brains, that doesn't make people equivalent to laptops. The key differences and key stuff to analyze are software not hardware. Again this is well known and I'm running into ignorance not counter-arguments.

Another problematic issue is the many people who, instead of trying to say how animals are partly like humans (elevating animals), instead argue to lower the value and importance of humans. While that approach isn't ruled out by logic, it's full of bad anti-human philosophy that has been refuted e.g. in DD's books (which these ppl don't understand, so the discussion becomes about trying to teach ideas to a hostile debater, not a curious learner, so it's hard to get anywhere).


curi at 6:59 PM on November 20, 2019 | #14443 | reply | quote

https://en.wikipedia.org/wiki/Chinese_room

This is, among other bad things, a denial of the hardware independence of computation. It does this by hypothetically supposing a type of computer which intuitively seems different than silicon CPUs, not with an argument directly about hardware independence, physics or computer science.


curi at 8:30 PM on November 20, 2019 | #14453 | reply | quote

Freeze pointed out that I already wrote this: Brains are Computers


curi at 11:16 PM on November 20, 2019 | #14468 | reply | quote

https://www.reddit.com/r/compsci/comments/e4urpw/the_programmers_challenge_to_animal_rights/

> As an activist for (other) animals rights I can tell you that there is zero need for this kind of arguments, the reasoning behind this activism is not rational but purely empathetic.

lol.

13 comments there now.

And past 50 comments at https://www.reddit.com/r/DebateAVegan/comments/e4ul7v/the_programmers_challenge_to_animal_rights/


curi at 11:19 AM on December 2, 2019 | #14652 | reply | quote

Some posts re animal rights.

https://thecriticalrationalist.weebly.com/on-the-inner-lives-of-animals.html

http://www.bretthall.org/humans-and-other-animals.html

Note: Brett Hall plagiarizes some of my and TCS's stuff. He tricks the weebly author into thinking arguments are "novel", which are due to DD and I, because he doesn't give credit for them. He does credit BoI for universality stuff, which is actually sneakier than if he credited nothing all post. He also gives other sources, e.g. Sam Harris links, again giving the false impression that he sources stuff as appropriate.

Brett's post also doesn't have a date on it.

This isn't Brett's first time plagiarizing me while also hating me simultaneously. Speaking to him about it won't help. I've tried that previously.


curi at 11:29 AM on December 2, 2019 | #14653 | reply | quote

https://www.reddit.com/r/DebateAVegan/comments/e4ul7v/the_programmers_challenge_to_animal_rights/f9k0f4d/

I liked this question (so I answered it) because I thought it was honest ignorance and willingness to ask about a basic point instead of saying a bunch of complicated stuff the person doesn't actually understand.


curi at 2:05 PM on December 3, 2019 | #14668 | reply | quote

#14668 My mistake, he wasn't honest or curious. He wasn't trying to learn or understand. He replied with severe goalpost moving and unacknowledged topic changing:

> I think my life has gotten just so much better when I started living it in accordance with my values like compassion and justice.

> Do you rememeber an event where you acted compassionately towards an animal? What did it feel like for you?


curi at 4:44 PM on December 3, 2019 | #14675 | reply | quote

Want to discuss this? Join my forum.

(Due to multi-year, sustained harassment from David Deutsch and his fans, commenting here requires an account. Accounts are not publicly available. Discussion info.)