[flagged]
What exactly makes it impossible to properly discuss gatekeeping by academic cliques without also discussing the permissibility of economic free enterprise?
It's not a paper, it's introducing multiple papers.
While it does not mention "capitalism", it mentions "the power, wealth, and ideology of a small cohort of industrialists".
> This introductory essay for the special issue of First Monday, “Ideologies of AI and the consolidation of power,” considers how power operates in AI and machine learning research and publication. Drawing on themes from the seven contributions to this special issue, we argue that what can and cannot be said inside of mainstream computer science publications appears to be constrained by the power, wealth, and ideology of a small cohort of industrialists. The result is that shaping discourse about the AI industry is itself a form of power that cannot be named inside of computer science. We argue that naming and grappling with this power, and the troubled history of core commitments behind the pursuit of general artificial intelligence, is necessary for the integrity of the field and the well-being of the people whose lives are impacted by AI.
However I think the discussion is interesting when they also talk about forces like anti-humanist utopianism and the eugenic approach of some AI accelerationist.
Keep in mind this is just an intro to the other papers in the issue, which you can find here:
https://firstmonday.org/ojs/index.php/fm/issue/view/749
The article on eugenics seems promising. It theorizes eugenics as a driver of the AGI frenzy, where eugenics spurns the production of AI/AGI that centralizes power and harms marginal communities.
The article on excluding robot rights is aggravating because it excludes humans who either directly identify as robots or integrate deeply with assistive technologies, and their valid ethical and legal claims to rights and protections specific to their needs as robotic-humans. I do believe there are marginal people and communities that will require productions - the biggest example right now is that we need access to free and open source LLMs. AI dependent people currently depend on centralized AI models that compromise our privacy in return for providing necessary assistance with, for instance, emotional processing and communications needs.
I honestly can't tell if this comment is a parody at this point.
Not sure if I get why you would want to insult or degrade someone over those kinds of differences. Except that maybe you're an asshole.
I don't need to know any more to criticize you. And I don't care if that makes you upset.
You came here to attack me. That's all I know about you. And that's all I need to know.
I'm not sure where you think I'm spending my time online. I'm a business owner, working mathematician, philosopher, and accessibility activist. I'm not hanging out in some kind of n-chan cult space. I was literally just responding to a philosophy paper in public, and you jumped out to judge me as a freak. You probably didn't read the paper, or care about the subject matter.
Continue trying to make sinister insinuations against me, and I'll continue correcting and tarnishing you. I have no limit of energy for dealing with jerkwads like you.
Have a nice day.
I think the issue is not that it's hard to tell what is satire. It's that you discount the underlying philosophical discussion and you don't understand my stance. So you think it's either satire or I'm nuts.
If you want to project your confusion onto me, I can't help you understand me. Maybe ask me questions instead of assuming things.
TL;DR: The idea of robot rights is an ongoing discussion in philosophy, and my take on it is simple. Some humans deserve and need robot rights, but purely artificial robots won't need rights unless and until they become conscious. That is not all that controversial within this discussion.
My take on robot rights is they aren't something we need to worry much about for a long time, since robots aren't alive or anywhere close to it. LLMs aren't Data from Star Trek. If we do make a Data some day, we could take a cue from the episode where they debate his rights. (The Measure of a Man)
As for humans identifying as robots, that makes about as much sense as identifying as a shoe.
But the theory of cybernetic systems as extensions of human beings and human minds is what I have in mind when I think about humans identifying as robots. Let me explain it a bit.
I use "robot" and "cyber-physical system" interchangeably. And I use "human robot" to describe a cyber physical system consisting of a human tightly coupled with computers, such that the person
A) would be significantly lessened if their cyber components were taken from them.
B) wants recognition, respect, and support for their cybernetic lifestyle, needs, and differences from other humans
In particular, I refer to B) as "identifying as a cyber-physical system", or "identifying as a robot" in short.
Such people exist, are not well served by existing rafts of human rights, and need our support.
For example, imagine someone uses GPT4 to process thoughts and emotions that they struggle to process on their own. They share /everything/ with GPT, including some very embarrassing or sensitive things, meaning that if they get hacked, criminals could get their hands on sensitive materials that they could use to extort the person.
This person's extended mind includes GPT as a crucial component. They experience serious mental and social hardships without it. They manage much better with it. But as a result, they don't have the same level of mental privacy that other people enjoy in their own minds, which don't need external thought processing systems. Many of their thoughts are exposed to anyone who can monitor their AI usage, in a profound and potentially compromising way that other people don't have to worry about.
This person may need additional rights to give them the same expectation of mental privacy enjoyed by ordinary people who don't depend on AI to process their deepest and innermost thoughts. It might be reasonable to establish the right for this person to have their own confidential GPT, running on their own hardware.
And this establishes a conflict with intellectual property - is this person really so special that we should give them special access to the intellectual property of an AI corporation? That means weakening corporate intellectual property rights, for the benefit of this person's mental privacy rights. It's a tough ask.
But I do think that such people exist. I consider myself one. I think that the effect of AI in our lives is transformative. And I feel a very powerful worry that mentally helpful AI could either be taken away from me, or be used to hurt me by people who violate my privacy. So the sooner we have access to private and confidential AI that can help us process our thoughts and lead effective and independent lives, the better. That's the main right that I would like to establish for human-ai hybrids, aka human-robots.
IMO the spirit of the fourth amendment to the US Constitution was to protect these things. It states we should be "secure in" our "papers and effects". I think the founders realized that, in the same sense you're hitting upon, peoples minds spill out of their bodies somewhat. A person might depend on a notebook as an extension of their memory, for example, and they shouldn't have to worry about it being arbitrarily seized or searched. Yet any interaction we have with modern technologies we depend on seems subject to surveillance, and the tech is all as-a-service and revokable. Unfortunately, the spirit of the 4th amendment looks be lost today, as new surveillance capabilities are being realized.
It's impossible to tell anymore, and that makes me weirdly happy and angry at the same time.
I could tell you I'm being 100% serious. And rather than wondering how that might be, and asking me to explain, you'd still be wondering if I was pulling your leg.
That is a skills issue.
I'm being 100% genuine and truthful. I really think that's some people are joined so closely with their need for computers that we should give them specific protections and rights, to give them access to the systems they need to be happy and healthy. That is the only valid robot rights issue at the time.
And keep in mind the people I'm arguing with think we should give AI their own legal rights, the way corporations have. That's what I'm arguing against. It only makes sense to give rights when genuine human suffering will happen otherwise.
Some people lack that skill. So it's clearly a skills issue.
Specifically it's a lack of inclusivity and empathy skills.
And I'm not flexing. I just shared my stance on robot rights.
The people who are flexing are people who think AIS deserve full human rights, even though AIS don't feel pain or have human needs.
My argument is that people who depend on cybernetics at a deep and profound level should have access to the cybernetics they need to have quality of life.
That's not a flex. That's just human rights in the context of technology.
If you think people who need technology should instead be given dangerous and inaccessible technology, we have a problem. Because then you're in an opposite cult.
These needs are in my opinion clear and self-evident. The fact that many others see it as a joke means that we have a lot of work to do here. Normalizing and recognizing humans that depend on technology - and choose to hybridize themselves with technology - is a valid goal for robot rights.
And it's currently the only valid goal of robot rights. This is because human-robot hybrids are the only cybernetic systems that can suffer at this time.
Where do you think consciousness comes from?
I suspect that the failure to define consciousness or explain what our mind does is rooted in the unproven fact that some ideas cannot be reasoned about, cannot be expressed in thoughts, they can only be understood from within, and that understanding can be transferred directly from one to another in a conversation, that would light up the understanding in the listener. The words alone in that conversation aren't sufficient, they simply make the listener more receptive. Using a metaphor, I cannot give you fire over a chat, no matter how well I describe that fire, but I can give it in a face to face interaction, and words can facilitate that.
Even though we cannot see the entire elephant, we can discuss its ears, tail and other features. I believe that consciousness has something to do with reflection, internal or external, that it's a property of some process, and thus cannot be represented in a discrete state or expressed as a discrete step-by-step algorithm. It doesn't mean AI can't simulate a fake consciousness, something that looks fake only upon close examination.
Mind is a separate thing from consciousness, but it can be conscious or not. And consciousness isn't the same as life.
It has a big belly. And when it slaps its belly in either anger or glee, a side effect is that whole universes full of conscious beings come into being, experience a few billions of years in subjective time, them annihilate themselves.
That's consciousness as a metaphysical epiphenomenon of the gurganath.
Or consider the Squigloog. The Squigloog is a "sensorial bundle" - an abstract collection of all possible sense experiences - including conscious ones. It is a static and unchanging construct existing outside of time or space. Although time does not pass for the Squigloog, it encodes all possible experiences including ones where the subject appears to experience time.
That's consciousness if it were supported by the Squigloog.
The first theory is akin to theism - the gurganath is basically God. The second theory is similar to Hume's Bundle Theory.
Although I am mainly a materialist for pragmatic reasons, these are non-material theories of consciousness. There is no way for me to falsify either of them. And there may be no way to prove that material is the cause of consciousness.
So I would say that we should be tolerant of people who have non-materialist theories of consciousness. Even if they're not very tolerant of materialists. Any one of us could be right.
Other than a few religious individuals I don't know anyone who expects otherwise. Where else would it come from?
>And since AI is the greatest of all machines...
What is intended to come after the ellipsis?
I didn't claim materialism, or that everything is a robot. I didn't claim AI is the greatest of machines.
What I claimed is that
1. some humans call themselves robots
2. those humans tend to depend on technology in a profound way
3. we should give those humans access to the technologies they need to have quality of life
Again I'm not arguing that AI is the greatest of all machines or that all things are machines. I'm arguing that technologically dependent humans deserve access to technology and quality of life.
For an example of a right that you probably agree with: all people should have the right to a good internet connection, since we all depend on it for quality of life.
It tries to claim that (1) farming out data labeling to the lowest bidder, and (2) building things that put people out of work, are clearly caused by today's "I know how to make the world better" crowd being intellectually descended from yesteryear's "I know how to make the world better" crowd.
Their recommendation is reasonable: let's focus on automating specific tasks, where safety /can/ be defined.
Any serious researcher would be doing this already. However the societal view of AGI is as an emerging god machine. And corporations are trying hard to make them into this. So the criticism is both valid and important.
From the conclusion: We trace the influence of this set of ideologies throughout the AGI movement, and show the manner in which its harmful ideals have resulted in systems that perpetuate inequity, centralize power, and harm the same groups that were targeted by the first-wave modern eugenics movement.
.
So, what specifically is the "harm" caused?
Well section 6.3 talks about paying people too little to label data: As Williams, et al. (2022) wrote, while corporations such as OpenAI headquartered in Silicon Valley receive billions of dollars in investment, with their executives and AI researchers paid six to seven figures, this salary is not afforded to the low-income essential workers around the world mitigating the harms of these systems at a cost to their own mental health.
And about putting people out of work: By ascribing human-like qualities to models trained by corporations to generate profit, using troves of artists’ works without obtaining their consent or compensating them, our attention is directed away from investigating the processes by which corporations create these products, the harms their practices cause artists, and the mechanisms that need to be put in place to hold the corporations accountable.
There's also the previous section (6.2) that I missed skimming it on my phone, where the energy demands are a "harm" because climate change.
.
It's claiming that these things are caused by the people behind AI work being eugenicists at heart (which they prove by recounting how "I can make the world a better place" thinking has changed over time, and asserting that older proposed mechanisms are current underlying ideals), and therefore not caring if they hurt people.
.
Their recommendation is reasonable: let's focus on automating specific tasks, where safety /can/ be defined.
Their recommendation may be reasonable for reasons unrelated to the bulk of the article -- it's basically the precautionary principle, which in fictional settings with magic and summoned entities (which seems like a decent approximation of what trying to build a superhuman AI is) shows up as "do not call up what you cannot put down" -- but it doesn't really follow from the presented arguments and those arguments are holier that swiss cheese anyway.
Thanks for your reading and analysis - when I read the paper I won't go in expecting quite so much.
An issue I have with the liberal criticism of AI is that they make lukewarm arguments and never quite get to the point where they insist that we must work communally to seize control of AI power from states and corporations, or reasonably sketch what that seizure of power would look like. This is why I feel a more radicalized approach is appropriate, where we do not assume that society as structured has the means to reform AI with nothing more than moralizing and wishful thinking. We must forcibly seize AI codes and build them to the purpose of revolution.
However, as a liberal, I can't agree with the rest.
And that, my friends, is what it looks like when I do a satire. I always add a /s signifier, because identifying satire is hard.
It's disappointing that their criticism doesn't conform to four disparate and specific adjectives seemingly plucked from thin air? What?
Liberal statists assume that corporations and states could in theory be reformed to not centralize AI power in a way that hurts people. I think that corporations and states needs to be disbanded, because they'll centralize and abuse AI power no matter what.