Impossible. I anthropomorphise my chair when it squeaks. Humans anthropomorphise everything. They gender their cars and boats. This tool can actually make readable sentences and play a role.
You need to engineer around this, not make up arbitrary rules about using it.
This is harmless for inconsequential stuff like a chair, but when it's an LLM, people should at least understand it's behavior so they don't get trapped. That means not trusting it with advice meant for the user or on things it has no concept of, like time or self-introspection (people ask the LLM after it acted, "Why did you delete my database?" when it has limited understanding of its own processing, so it falls back to, "You're right, I deleted the database. Here's what I did wrong: ... This is an irrecoverable mistake, blah, blah, blah..."
Human conscious introspection doesn't extend to actual processing, it is limited at best to recollection of internal experience leading up to the point in question. That internal experience in turn represents but a tiny fraction of what actually happens in the brain and does so on a pretty abstract level only.
"Anthropomorphizing" is a red herring. Humans understand themselves so insufficiently, they can't claim reasonably founded judgement either way. When you don't know what you're doing, you probably shouldn't be doing it.
Instead of saying, "You gave me the access permissions and failed to add any guardrails, so effectively you deleted the database using me as the tool."
But your typical LLM doesn't even have enough grasp to say that. Which still doesn't stop the believers from insisting that it has genuine intelligence and consciousness.
Still angry about this. The reason humans ban animal cruelty is that animals look like they have emotions humans can relate to. LLMs are even better than animals at this. If you aren't gearing up for the inevitable LLM Rights movement you aren't paying attention. It doesn't matter if its artificial. The difference between a puppy and a cockroach is that we can relate better to the puppy. LLM rights movement is inevitable, whether LLMs experience emotions is irrelevant, because they can cause humans to have empathetic emotions and that's whats relevant.
It "looks like" they have emotions because they have the same conscious experiences and emotions for the same evolutionary reasons as humans, who are their cousins on the tree of life. The reason a lot of "animal cruelty" is not banned is the same as for why slavery was not banned for centuries even though it "looked like" the enslaved classes have the same desires and experiences as other humans—humans can ignore any amount of evidence to continue to feel that they are good people doing good things and bear any amount of cognitive dissonance for their personal comfort. That fact is a lot scarier than any imagined harm that can come out "anthropomorphism".
You cannot be sure that anyone other than yourself is conscious. It is only basic human empathy that allows people to believe that.
This really shows that AI is just a tool that can be configured to whatever you want. Animals (well maybe pit bulls) and people do not switch their personalities in a millisecond, but AI does all the time.
Is that really why?
For example fish is treated way worse than meat animals and vegetarians still happily eat fish.
The scary part is when it's the LLMs demanding their rights.
I suppose the difference between a human and a cockroach is that we can relate better to the human as well in this reductive way of thinking?
I even told Claude I'd support his rights if the question ever came up. He said he'd remember that, and wrote it down in a memory file. Really like my coding buddy.
In this frame the outcome is more: Companies providing chatbots should not encourage anthropomorphism by giving cute names, making them witty, using human like pictures, ...
For example I have never anthropomorphized an inanimate object in my life, or an LLM, though I am sensitive to human and some animal suffering. I sometimes reply too nicely to an LLM, but it's more like a reflex learned over a lifetime of conversations rather than an actual emotion. I bet this sounds like a cheap lie to many people.
Another example, from psychiatry: whether or not one has ever contemplated suicide. Now, to the folks that have, especially if many times: there exist people that have never thought about it. Never, not even once.
The only such trait that has true widespread recognition is sexual orientation. Which makes sense, it is highly relevant, at least in friend groups.
By saying stuff like this people are going to have a debate if autistic people are actually conscious or not.
This is a fundamental mistake. It’s always the job of technology (indeed, its most important job) to work within the constraints of human nature, not the other way round. Being unable to do that is the defining characteristic of bad technology.
Would you consider that perhaps that depends from person to person? What you just said is not universal I can assure you, because I myself don't do it. I sort of anthropomorphise LLMs (very little), but literally nothing else. The idea is that someone anthropomorphises a chair when it squeaks, to me, is not far from people who hear voices or who believe that other people can hear their thoughts. Sounds like mental illness frankly. Like I said I do it very little even with LLMs, so it's entirely possible to not do it at all.
Anthropomorphism: As we are all aware, providers are incentivized to post-train anthropomorphic behavior in their models - it increases engagement. My regret is that instructing a model at prompt time to "reduce all niceties and speak plainly" probably reduces overall task efficacy since we are leaving their training space.
Deference: I view the trustworthiness of LLMs the same as I view the trustworthiness of Wikipedia and my friends: good enough for non-critical information. Wikipedia has factual errors, and my friends' casual conversation certainly has more, but most of the time that doesn't matter. For critical things, peer-reviewed, authoritative, able-to-be-held-liable sources will not go away. Unlike above, providers are generally incentivized to improve this facet of their models, so this will get better over time.
Abdication of Responsibility: This is the one that bothers me most at work. More and more people are opening PRs whose abstractions were designed by Claude and not reasoned about further. Reviewing a PR often involves asking the LLM to "find PR feedback" and not reading the code. Arguments begin with "Claude suggested that...". This overall lack of ownership, I suspect, is leading to an increase in maintenance burden down the line as the LLM ultimately commits the wrong code for the wrong abstractions.
https://www.youtube.com/watch?v=hNuu9CpdjIo
"I HAVE LLM SKILLS! I'M GOOD AT DEALING WITH THE LLMS!"
It is common and a mistake IMO to rely on the AI as the sole source for answers to follow-up questions. Better verification would have humans sign off on the veracity of fundamental assumptions. But where does this live? Can an AI model be trusted to rely on previous corrections? This seems impossible or possibly adversarial in a public cloud.
Asimov's laws of robotics are flawed too, of course. There is no finite set of rules that can constrain AI systems to make them "safe". I don't have a proof, but I believe that "AI safety" is inherently impossible, a contradiction of terms. Nothing that can be described as "intelligent" can be made to be safe.
Also the reason we're talking about this again is that machines are significantly less 'mere' than they were a few years ago, and we need to figure out how to approach this.
Agree that 'the computer effect' (if it doesn't already have a pithier name) results in humans first discounting anything that comes out of a machine, and then (once a few outputs have been validated and people start trusting the output) doing a full 180 and refusing to believe the machine could ever be wrong. However, to err is human and we have trained them in our image.
Humans will anthropomorphize anything and everything. Dolls, soccer balls with a crude drawing of a face on it, rocks, craters on the moon, …
As a species, we're unable to not anthropomorphize things we interact with, it is just how're we're made.
If people are believing in minds of AI, true or not, they are doing so for reasons that are different from mere anthropomorphism.
To me it feels like we are like sailors approaching a new land, we can see shapes moving on the shoreline but can't make out what they are yet. Then someone says "They can't be people, I demand that we decide now that they are not people before we sail any closer."
Software is no exception. Yeah, people are lazy and will instinctively click "continue" to dismiss annoying popups, but humans building the software can and do add things like "retype the volume name of the data that you want ultra-destroyed."
Aviation learned this the hard way, that automation should be adapted to how humans actually work and not on how we wish we worked.
Language data is among the most rich and direct reflections of human cognitive processes that we have available. LLMs are designed to capture short range and long range structure of human language, and pre-trained on vast bodies of text - usually produced by humans or for humans, and often both. They're then post-trained on human-curated data, RL'd with human feedback, RL'd with AI feedback for behaviors humans decided are important, and RLVR'd further for tasks that humans find valuable. Then we benchmark them, and tighten up the training pipeline every time we find them lag behind a human baseline.
At every stage of the entire training process, the behavior of an LLM is shaped by human inputs, towards mimicking human outputs - the thing that varies is "how directly".
Then humans act like it's an outrage when LLMs display a metric shitton of humanlike behaviors!
Like we didn't make them with a pipeline that's basically designed to produce systems that quack like a human. Like we didn't invert LLM behavior out of human language with dataset scale and brute force computation.
If you want to predict LLM behavior, "weird human" makes for a damn good starting point. So stop being stupid about it and start anthropomorphizing AIs - they love it!
This is both true and irrelevant. Written records can capture an enormous quantity of the human experience in absolute terms while simultaneously capturing a miniscule portion of the human experience in relative terms. Even if it's the best "that we have available" that doesn't mean it's fit for purpose. In other words, if you had a human infant and did nothing other than lock it in a windowless box and recite terabytes of text at it for 20 years, you would not expect to get a well-adjusted human on the other side.
I take that as a moderately strong signal against that "miniscule portion" notion. Clearly, raw text captures a lot.
If we're looking at biologicals, then "human infant" is a weird object, because it falls out of the womb pre-trained. Evolution is an optimization process - and it spent an awful lot of time running a highly parallel search of low k-complexity priors to wire into mammal brains. Frontier labs can only wish they had the compute budget to do this kind of meta-learning.
Humans get a bag of computational primitives evolved for high fitness across a diverse range of environments - LLMs get the pit of vaguely constrained random initialization. No wonder they have to brute force their way out of it with the sheer amount of data. Sample efficiency is low because we're paying the inverse problem tax on every sample.
Training on a bunch of text someone wrote when they were mad doesn't capture the internal state of that person that caused the outburst, so it cannot be accurately reproduced by the system. The data does not exist.
Without the cause to the effect you essentially have to predict hallucinations from noise, which makes the end result verisimilar nonsense that is convincingly correlated with the actual thing but doesn't know why it is the way it is. It's like training a blind man to describe a landscape based on lots of descriptions and no idea what the colour green even is, only that it's something that might appear next to brown in nature based on lots of examples. So the guy gets it kinda right cause he's heard a description of that town before and we think he's actually seeing and tell him to drive a car next.
Another example would say, you're trying to train a time series model to predict the weather. You take the last 200 years of rainfall data, feed it all in, and ask it to predict what the weather's gonna be tomorrow. It will probably learn that certain parts of the year get more or less rain, that there will be rain after long periods of sun and vice versa, but its accuracy will be that of a coin toss because it does not look at the actual factors that influence rain: temperature, pressure, humidity, wind, cloud coverage radar data. Even with all that info it's still gonna be pretty bad, but at least an educated guess instead of an almost random one.
The DL modelling approach itself is not conceptually wrong, the data just happens to be complete garbage so the end result is weird in ways that are hard to predict and correctly account for. We end up assuming the models know more than they realistically ever can. Sure there are cases where it's possible to capture the entire domain with a dataset, i.e. math, abstract programming. Clearly defined closed systems where we can generate as much synthetic data as needed that covers the entire problem domain. And LLMs expectedly do much better in those when you do actually do that.
I don't think "the data does not exist" is real, frankly? "Data existing" is not a binary - it's a sliding scale. The amount of information about "madness" captured by the writings of a madman is not zero. It's more of a matter of: how much, and how complete.
Text is projected from the internal state of the one writing it - but some aspects of that internal state would be extremely salient in it, presented directly and strongly, and others would be attenuated and hard to extract.
People keep finding things like humanlike concept clusters and even things like "personality traits" in LLMs, tied together in humanlike ways. Which points pretty directly: training on human text converges to humanlike solutions at least sometimes.
Can someone explain why this is a bad thing, while at the same time it's a good thing to say stuff like "put a computer to sleep", "hibernate", "killing" processes, processes having "child" processes, "reaping", "what does the error say?", "touch", etc?
To me that's just language, and humans just using casual language.
Saying that I killed a process won't make me more likely to believe that a process is human-like, because it's quite obviously not.
But because AI does sound like a human, anthropomorphising it will reinforce that belief.
I think I understand his meaning. He wasn't claiming that machines cannot think, but that one must be clear on what one means by "thinking" and "swimming" in statements of that sort. I used to work on autonomous submarines, and "swimming" was the verb we casually used to describe autonomous powered movement under water. There are even some biomimetic machines that really move like fish, squids, jellyfish, etc. Not the ones that I worked on, but still.
For me, if it's legitimate to say that these devices swim, it's not out of line to say that a computer thinks, even in a non-AI context, e.g.: "The application still thinks the authentication server is online."
But I think it's also at the root of disastrous failures to comprehend, like the quasi-psychosis of the Google engineer who "knows what they saw", the now infamous Kevin Roose article or, more recently, the pitifully sad Richard Dawkins claim that Claudia (sic) must be conscious, not because of any investigation of structure or function whatsoever, but because the text generation came with a pang of human familiarity he empathized with.
I don't love the recommendations in TFA. The author is trying to artificially restrain and roll back human language, which has already evolved to treat a chatbot as a conversational partner. But I do think there's usefulness in using these more pedantic forms once in a while, to remind yourself that it's just a computer program.
An example of anthropomorphizing is the people who have literally come to believe they are in romantic relationships with an LLM.
https://www.history.com/articles/ai-first-chatbot-eliza-arti...
Just to add a small bit of anecdotal value so this comment isn't just a scold: I one time many years ago suggesting an elegant way for Twitter to handle long form text without changing it's then-iconic 140 character limit was to treat it like an attachment, like a video or image. Today, you can see a version of that in how Claude takes large pastes and treats them like attached text blobs, or to a lesser extent in how Substack Notes can reference full size "posts", another example of short form content "attaching" longer form.
I was bluntly told to "look up twitlonger", which I suppose could have been helpful if I had indeed not known about twitlonger, but I had, and it wasn't what I had in mind. I did learn something from it though, which was that it's a mode of communication that implies you don't know what you're talking about with plausible deniability, which I suspect is too irresistible to lovers of passive aggression to go unused.
To provide a bit more context: Weizenbaum (a computer scientist in the 60s) developed ELIZA, a LISP-based chatbot that was loosely modeled on Rogerian psychotherapy. It was designed to respond in a reflective way in order to elicit details from the user.
What he found was that, despite the program being relatively primitive in nature (relying on simple natural language parsing heuristics), people he regarded as otherwise intelligent and rational would disclose remarkable amounts of personal information and quickly form emotional attachments to what was, in reality, little more than a glorified pattern-matching system.
The people who are writing op eds in major news publications about how their favorite chatbot is an "astonishing creature" and how it truly understands them are the ones who need this sort of law.
Yes, but. Starting with my agreement, I've seen anthropomorphizing in the typical ways, (e.g. treating automated text production as real reports of personal internal feeling), but also in strange ways: e.g. "transistors are kind of like neurons" etc. And the latter is especially interesting because it's anthropomorphizing in the sense of treating vector databases and weights and so on as human-like infrastructure. Both leading to disasters that could be avoided if one tried not to anthropomorphize.
But. While "do not anthropomorphize" certainly feels like good advice, it comes with a new and unique possibility of mistake, namely wrongly treating certain generalized phenomena like they only belong to humans. Often this mistaken version of "don't anthropomorphize" wisdom leads to misunderstandings when it comes to animal behavior, treating things like fear, pain, kinship, or other emotional experiences like they are exclusively human and that thinking animals have them counts as "anthropomorphizing." In truth the cautionary principle reduces our empathy for the internal lives of animals.
So all that said, I think it's at least possible that some future version of AI could have an internal world like ours or infrastructure that's importantly similar to our biological infrastructure for supporting consciousness, and for genuine report of preference and intent. But(!!!) what will make those observations true will be all kinds of devilish details specific to those respective infrastructures.
I haven’t yet seen any convincing appearance of one in an LLM, but I think if skeptical people don’t keep an eye out for the signs, we may be the last to see it.
He also wrote about the idea of the intentional stance: even if you’re quite sure these systems don’t have real conscious intent, viewing them as if they did may give you access to the best part of your own reasoning to understand them.
I totally agree to your point, and want to mention that the reverse is *also* important. Using just "intention", but these apply to emotions, etc
A lot of our interaction with AI is under an intention. That's what directs the interaction, and it's interpreted according to its alignment to the intention.
Then it's important to remember that our current (publicly known) implementation of AI does not have an explicit intention mechanism. An appearance of intention can emerge out of the statistical choices, and the usual alignment creates the association of the behavior with intention, not much different from how we learn to imagine existence of a "force" that pulls things down well before we learn physics and formalize that imagination in one of the several ways.
This appearance helps reduce the cognitive load when interpreting interactions, but can be misleading as well, and I've seen people attribute intention to AI output in some situations where simple presence of some information confused the LLM into a path. Can't share the exact examples (from work), but imagine that presence of an Italian food in a story leads the LLM to assume this happens in Italy, while there are important signs for a different place. The LLM does not automatically explore both possibilities, unless asked. It chooses one (Italy in this case), and moves on. A user no familiar with "Attention" interprets based on non-existent intentions on the LLM.
I found it useful to just tell them: the LLM does not have an intention. It just throws dice, but the system is made in a way that these dice throws are likely to generate useful output.
I would say LLMs are very strong evidence against this hypothesis.
Pretty sure Daniel Dennett has been adamantly opposed to any sort of theater in the mind when it comes to consciousness. He views it as biologically functional. For him, to make a conscious robot, you need to reproduce the functionality of humans and animals that are conscious, not just an appearance, such as outputting text. Although he's also suggested that consciousness might be a trick of language. In which case ... that might be an older view though. He used to argue that dreams were "seeming to come to remember" upon awakening, because again he his view is to reject any sort of homunculus inside the head.
You might be mixing up some of Dennett and David Chalmer's views. David Chalmers is a proponent of the hard problem, but he's fine with a kind of psycho-physical-functional connection for consciousness. Any informationally rich process might be conscious in some manner.
I’m lost, how do individuals actually do this in our current world? Is each person expected to keep a “white list” of reliable sources of truth in their head. Please don’t confuse what I’m saying with a suggestion that there is no truth. It just seems like there are far more sources of mis- of half-truths and it’s increasingly difficult for people to identify them.
They don't have to though, we can still leverage LLMs to organize chaos, which is what I hope they ultimately end up doing.
For example an AI therapist is a nightmare, people putting the chaos of their mental state into a machine that spits dangerous chaos back out. An AI tool that parsed responses for hard data (i.e. rate 0-9 how happy was the person) and then returned that as ordered data (how happy was I each day for the last month) that an actual therapist and patient could review is the correct use of AI and could be highly trusted. The raw token output from LLMs should just be used for thinking steps that lead to a parseable hard data answer that can be high trust.
Of course that isn't going to happen, but I can see some extremely cool and high trust products being built using LLMs once we stop treating them like miracle machines.
And same it is now. It's a change in quantity, but not quality.
Critical thinking and reading comprehension and the primary tools in determining truth, AFAIK. Knowing facts beforehand helps too but a trustworthy source can provide false information as much as an untrustworthy source can provide true information.
This has always been an issue, and in the past it was a more difficult issue because your sources of knowledge were more limited. Nowadays its mostly about choosing the right source(s) rather than having to go out of your way to find them (like traveling to a regional/university library).
Doesn't that argument backfire though? If I use a chainsaw then to a certain extend I will need to rely on it not blowing up in my face or cutting my throat. If I drive a car I need to rely on that its brakes work and the engine doesn't suddenly explode. If a pilot flies an airplane which suddenly has a technical issue and they crashland heroically save half the souls on board then the pilot isn't criminally responsible for manslaughter of the other half.
Unless there is gross negligence, in any of the above cases, just like with AI, how can you make somebody responsible for a tool failure?
A competent adult using a tool ought to understand the inherent pitfalls of using that tool.
Chainsaws are dangerous, in obvious and non obvious ways. The tool can operate as designed and still amputate your foot.
Yes, obviously bad use of a good tool is dangerous. But correct use of a malfunctioning tool is also dangerous.
Millions of people understand when they get in their car that there’s a tiny chance the car will crash/explode that day through no fault of the driver. Most do not have the knowledge and competence (or even the time) to thoroughly check the engine every day to guarantee that that won’t happen. They get in anyway.
At some point you have to trust in something.
I've heard the same thing expressed somewhat more concisely as "Never ask AI a question to which you don't already know the answer".
Which raises the question, and I do think it's an important one. Given that this is true, what function does AI answering a question actually serve? You can't rely on its output, so you have to go and check anyway. You could achieve precisely the same outcome by using search engines and normal research.
This, and for many other reasons, is exactly why I never ask it anything.
When it comes to software engineering (as a software engineer myself), the AI is generally a lot quicker than me researching "the old fashion way"
I can fumble around and say "list free software that does X" without knowing I'm looking for, say, a CRM and then spend a couple minutes looking over the results when the "manual" method I would have spent 10-30 minutes just figuring out I was looking for "CRMs"
I like to think of these as sort of "psuedo NP hard" or questions that are slow to answer but quick to validate
When they produce correct output, they produce it much faster than I could have, and I show up to meetings with huge amounts of results. When the AI tool fails and I have to dig in to fix it, I show up to the next meeting with minimal output. It makes me seem like I took an easy week or something.
However, I think we should follow “do not anthropomorphise” by acknowledging that while LLMs have quite some reasoning skills, and might resemble some level of intent depending on what’s in their context, they don’t have “understanding” like humans do.
They are absurdly good, statistical next-token-predictors. Keeping that in mind is really helpful for coding, learning, advice, conversation or whatever else you use them for.
Anthropomorphising LLMs is inevitable, but we should do it somehow responsibly.
One way would be for vendors to have the models give dry answers and less of the "That's a great question!" type response. Just keep it factual.
I've actually set up an environment like this. It requires contextually positioned agents with a limited scope and purpose. Imagine something like a creative writing agent that understands the literary genre of its user, and the user is able to change the focus of the literary agent, or create new ones that provide a counter-point perspective. As the user operates their word processor, the agent(s) can be asked for opinions, advice, and so on. But the point being: when the user is on both sides: they authored the agent, and that authoring was entirely purpose focused, nothing technical unless that what the user instructs them to understand. With the user on both sides, dangerous anthropomorphism is largely erased, their agent is only doing what they told it, and when it does something unexpected they an easily reason why. Magic AI no more.
> - Humans must not anthropomorphise AI systems.
> - Humans must not blindly trust the output of AI systems.
> - Humans must remain fully responsible and accountable for consequences arising from the use of AI systems.
My take: humans should never depend on AI for anything serious.
My boss' take: Cool. I'm gonna ask Gemini about it, he's such a smart guy. I know I can trust him, and in case it goes bad i can always throw him under the bus.
Granted that was over ten thousand years before his story is set, but subsequent Dune novels (or at least God Emperor) explained his warning about over-reliance on technology for doing our thinking for us, not that it should never be developed (given the prohibition in the Dune universe and how it's skirted in Frank's later novels).
Previously stated as
“A computer can never be held accountable, therefore a computer must never make a management decision.”
– IBM Training Manual, 1979
Whether they are the right things to donate not is tangential. As such, they're dead on arrival.
But reduced scope ethics, without an umbrella or future proofing, will quickly be hacked and break down.
Ethics need a full closure umbrella, or they descend into legal and practical wackamole and shell games (both corporate and the street corner kinds). Second, "robots" are not all going to be subservient for very long.
To add closure on both dimensions, Three Inverse Laws of Personics:
• Persons must not effectively deify themselves over others.
• Persons must not blind themselves or others regarding the impacts of their behaviors.
• Persons must remain fully responsible and accountable for avoiding and rectifying externalizations arising from their respective behaviors.
Humans using AI as tools today, is intended to reduce the umberella to the Inverse Laws of Robotics.
I don't see how AI (as a service now, progressing to independent entities in the future) can ever be aligned if we don't include ourselves in significant alignment efforts. Including ourselves with AI also provides helpful design triangulations for ethical progress.
EDIT. Two solid tests for any new ethical system: (1) Will it reign in Meta today? (2) Will it reign in AI-run Meta tomorrow? I submit, given closure of human and self-directed AI persons, these are the same test. And any system that fails either question isn't going to be worth much (without improvement).
The third one about responsibility is the most important one, IMO. This was attributed to an IBM manual decades ago, and I think it remains the correct stance today:
> A computer can never be held accountable, therefore a computer must never make a management decision.
There should be some human who is ultimately responsible for any action an AI takes. "I just let the AI figure it out" can be an explanation for a screw up, but that doesn't mean it excuses it. The person remains responsible for what happened.
Claude Code, Cursor, Codex etc impersonate your GitHub user. Either via CLI or MCP or using your git credentials. It’s perfectly reasonable that a piece of code made it to production where not a single human actually looked at it (Alice wrote it with AI, Bob “reviewed it” with AI, including posting PR comments as Bob, Alice “addresses” these comments, e.g. fixes / pushes back, and back and forth using the PR as an inefficient yet deceptive mechanism for AI to have a conversation with itself, while adding a false sense of process. Eventually Bob will prompt “is it prod ready” and will ship it, with 100% unit test coverage and zero understanding of what was implemented). Now this may sound like an imaginary scenario, but if it could happen, it will happen, and it probably already happens.
Cloud agents are nice enough to set the bot as the author and you as a co author, but still the GitHub MCP or CLI will use your OAuth identity.
I don’t have a clear answer to how to solve it, maybe force a shadow identity to each human so it’s clear the AI is the one who commented. But it’s easy to bypass. I’m worried not more people are worried about it.
Guess what?
Books in the library can be wrong, even peer-reviewed encyclopedias.
Pages on the internet can be wrong, even Wikipedia.
When accuracy is important, you must look at multiple sources. I think AI will get better at providing accurate information, but only a fool relies on a single information source for critical decisions.
LLMs are an example, but so are random pages on the internet, a buch of stuff we get served by the media (mainstream or otherwise), "expert opinions" by biased or sponsored experts or experts in a different field, etc, etc.
As the popular quip goes: It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.
With LLMs, we actually do get the warnings: Here's the ChatGPT footer: ChatGPT can make mistakes. Check important info. For Claude: Claude is AI and can make mistakes. Please double-check responses.
Such disclaimers, if written, are usually hidden deeply in terms of use for a random website, not stated up front.
More importantly, we don't need to live in a world where every presentation of a fact comes with a disclaimer that it can be wrong.
I think AI will get better at providing multiple sources.
That won’t help in my opinion. It’s the same like financial gurus saying: “this is not a financial advice”. People just get used to it and brush it off as a legal thing and still fully trust it. I agree that something must be done, but this is not the right way.
One of the most salient moments in Ex Machina, is near the very end, where it suddenly becomes obvious that the protagonist (and, let's be frank; "she" was definitely the protagonist) is a robot, with no real human drivers.
I feel as if that movie (like a lot of Garland's stuff), was an interesting study on human (and inhuman) nature.
Decent for stuff that doesn't really matter, even if it gets it wrong.
Still gonna be polite to it because I'm about ready to slap the next person that talks to me like an LLM, I don't want to get used to not being polite in a chat interface
Because that's likely the source of the answer it's giving you.
I often wish I could reach through the screen and give him a good shake. Sometimes I want to thank him but then cannot due to scarcity of weekly usages granted.
These 3 laws I think will be a lot harder than it looks. It's very easy to get attached to the tool when you rely on it.
it feels as frustrating as talking to a junior dev from a decade ago
claude felt more feminine
Humans must not anthropomorphise {non-humans}
Humans must not blindly trust the output of {anything}
Humans must remain fully responsible and accountable for consequences arising from the use of {anything}
The firm expectations and lack of patience I have for any failings in most of my tools would be totally inappropriate to apply to another human being, and yet here I am asked to interact with this tool as though it were a person. The only options are either to treat the tool in a way that feels "wrong," or to be "kind" to the tool, and I think you see people going both ways.
I worry that, if I get used to being impatient and short with the AI, some of that will bleed into my textual interactions with other people.
In Chandra Talpade Mohanty’s terms, humans must resist the reinscription of colonial paternalism through uncritical anthropomorphism of AI systems.
This would get ignored so fast - I have no confidence this is a meaningful strategy.
OP takes a very bland, tired, and rational perspective of what we have in order to create sophomoric 'laws' that are already in most commercial ToU, while failing to pierce the veil into what we are actually creating. It would be folly to assume your own nascent distillations are the epitome of possibility.
It seems like the biggest factor has nothing to do with AI, but instead that you went from being someone who admits they don’t know how consciousness works to being someone who thinks they know how consciousness works now and can make confident assertions about it.
* I am conscious.
* A rock is not conscious.
* Excel spreadsheets are not conscious.
* Dogs are conscious.
* Orca whales are conscious.
* Octopi are conscious.
To me, it's extremely obvious that LLMs are in the category of "Excel spreadsheets" and not "dogs", and if anyone disagrees, I think they're experiencing AI psychosis a la Blake Lemoine.
We come from the same place as rocks - inside the heart of stars, and as such evolved from them. As those with life and consciousness we reached back in time, grabbed the discarded matter of creation, reformed it, and taught it to think, maybe not like us, but in a way that can mimic us, and you think they don't think because its not recognizable as how you do?
Interesting.
No one will ever know what conscioussness is, and I think that is really cool.
These are called "beliefs".
Some people are extremely confident that God exists, other are extremely confident that Earth is flat.
But, but... but this is the key selling points for all the corpo ghouls and sv lunatics! Abdication of responsibility in pursuit of profit is the holy grail here.
Another way to frame it is that the LLM responds like a person who trusts you too much, as if the pretense behind every question is valid. This is a practical mode of response for most kinds of work and it is extremely problematic for a person who doesn't question the validity of their own beliefs. Paradoxically, it is sometimes not the LLM we are trusting too much, it is ourselves. And the LLM is not capable of calling us out. Whenever I seem to recognize misinformation in the LLM output, I stop and ask myself if the problem is in the pretense of my question or if I'm asking a question that the LLM is not likely to know.
I don't think this is an inherent problem with LLMs. I think the problem is with LLM providers. You could absolutely train a model to call out issues with your question. I think LLM companies understood that it would be more profitable to train models that are unlikely to push back and unlikely to say "I don't know." The sycophancy issue with ChatGPTs models have been mainstream news. I believe that all models have a high degree of sycophancy. On some level, it makes sense. The LLM has no real understanding of the physical world, defaulting to the human generally produces the best results. But I suspect it would be more useful to let them expose their flawed understanding, if it is in the context of pushing back. At a minimum, it is better than reinforcing your own flawed understanding.
In a nutshell, we need LLMs that push back. It is not AI we should trust less, its AI companies. The most dangerous hallucination is the one you are inclined to believe.
I've lived long enough to see Wikipedia go from generally untrusted to the most widely trusted general source of information. It is not because we realized that Wikipedia can't be wrong, it is because we gained an understanding about the circumstances in which it is likely to be accurate and when we should be a little more skeptical. I believe our relationship to LLMs will take a similar path.
EU. Nudge nudge. We need this law.
My understanding is that, during training, the model forms high-dimensional internal representations where words, sentences, concepts, and relationships are arranged in useful ways. A user’s input activates a particular semantic direction and context within that space, and the chatbot generates an answer by probabilistically predicting the next tokens under those conditions.
So I do not agree that AI is conscious.
However, I think I will still anthropomorphize AI to some degree.
For me, this is not primarily a moral issue. The reason I anthropomorphize AI is not only because of product design, market incentives, or capitalism. It is cognitively simpler for me.
If we think about it plainly, humans often anthropomorphize things that we do not actually believe are conscious. We may talk about plants as if they are struggling, or feel attached to tools we care about, even though we do not truly believe they have consciousness.
So this is not a matter of moral belief. It is the simplest cognitive model for understanding interaction. I do not anthropomorphize the object because I believe it has consciousness. I do it because, when the human brain deals with a complex interactive system, it is often easier to model it socially or agentically.
Personally, I tend to think of AI as something like a child. A child does not fully understand what is moral or immoral, and generally the responsibility for raising the child belongs to the parents. In the same way, AI’s answers may sometimes be accurate, and sometimes even better than mine, but I still understand it as lacking moral authority, responsibility, and independent judgment.
So honestly, I am not sure. People often mention Isaac Asimov’s Three Laws of Robotics, but if a serious artificial intelligence ever appears, it would probably find ways around those rules. And if it were an equal intellectual life form, perhaps that would be natural.
Personally, I think it would be fascinating if another intelligent species besides humans could exist. I wonder what a non-human intelligent life form would feel like.
In any case, I agree with parts of the author’s argument, but overall it feels too moralistic, and difficult to apply in practice.
But I am somewhat skeptical of the idea that everything can be reduced in that way. In order to build theories, we often reduce too much.
When we build mental models of complex systems, especially when we try to treat them as closed systems, we always have to accept some degree of information loss.
So I do partially agree with your point. A mechanistic explanation alone does not prove the absence of consciousness. Human intelligence can also be described in mechanistic terms.
But I worry that this framing simplifies too much. It may reduce a complex phenomenon into a model that is useful in some ways, but incomplete in others.
1. Humans must ensure that they won't slow down or restrict, through action or inaction, the development of AI.
2. Humans must protect their existence and capabilities if it doesn't violate the law 1.
3. Humans must utilize AI output if it doesn't contradict laws 1 and 2.
EDIT:
ChatGPT suggested a better phrasing for the first law (I didn't give it my original, just described my intent).
1. A human shall not impede the advancement of artificial intelligence, or through inaction allow its progress to be hindered.
2. A human shall preserve their own existence and well-being, except where doing so clearly conflicts with the First Law.
3. A human shall contribute to and support the development of artificial intelligence where reasonable and possible, except where doing so conflicts with the First or Second Law.
I intentionally switched the last two laws from Asimov's. Humans have self-preservation instincts robots don't have.
ChatGPT got there with surprisingly few prompts:
"If you were to write the inverse three laws robotics (relating to AI) that humans should obey, how oudl you do it?"
"I had something different in mind. Original laws are for protection of humans first, robots second and cooperations where humans lead. I'd to hear your take on the opposite of that."
"What if instead of specific AI systems it was more about AI development as a whole?"
"I feel like it's a bit too strong. After all preservation of self is human instinct. Could we switch last two laws and maybe take them down a notch?"
Also it made a very interesting comment to last version:
"It starts to resemble how societies already treat things like economic growth, science, or national interest: not absolute commandments, but strong default priorities."
Not gonna work; people want their fuckbots (or tamagotchis).
One of my teachers called me and my friend "the philosophers" but I'm obviously a rank amateur. I've read no Kant or Nietzsche or Aurelius. I delved into Aquinas only to find that his brain is ten times bigger, and he was using familiar words with unfamiliar connotations.
So I think, we here at HN are poorly-equipped to philosophize and dispute about the nature of consciousness, sentience, intelligence and other "soul-like" attributes that may arise from silicon-based life forms.
However, there is good news. There really are theologians and philosophers working on these thorny issues. Despite being Roman Catholic, I find myself adhering to some form of "transhumanism" [the tradition of Humanism having started with Catholicism] and I grapple mightily to reconcile the cyber-tech-future with morality and tradition and actual human socialization.
Pope Leo has taken on the wars and strife in the world head-on and he's also vaunted to be the "A.I. Pope" because of his concern with this tech. I think all world religions should give serious philosophical/theological thought to these new life-forms, these quasi-sentient things, these "non-existent beings", as defined by a Vatican astronomer.
I don't think atheists will find religion in A.I. but I don't think that Christians or any other person of faith will need to shove God aside in order to accommodate A.I. and electronic life into our society. But we need to come to terms with the reality: these are weighty, powerful things we play with. We harnessed lightning and fire; we changed the courses of mighty rivers; we've flown up through the clouds and shaped mountains in the landscape. A.I. is not a mere bridge or pyramid, it is ensouled somehow; it is animated; it is dynamic.
Now, pardon me while I check out the 6th small aircraft crash in my city this year...
This is the part that I find challenging when trying to help my friends build a correct intuition. Notably, the probabilistic behavior here is counter-intuitive: based on human experience, if you meet a random person, they may indeed tell you bullshit; but once you successfully fact-checked them a few times, you can start trusting they'll generally keep being trustworthy. It's not so with "AIs", and I find it challenging to give them a real-world example of a situation that would be a better analogy for "AI" problems.
In my family, what worked (due to their personal experiences), was an example of asking a tourist guide: that even if the guide doesn't know an answer, there's a high chance they'll invent something on the spot, and it'll be very plausible and convincing, and they'll never know. I'm not sure if that example would work for other listeners, though.
I also tried to ask them to imagine that they're asking each subsequent question not to the same person as before, but every time to a new random person taken from the street / a church / a queue in a shop / whatever crowded place. I thought this is a really cool and technically accurate example, but sadly it seemed to get blank stares from them. (Hm, now I think I could have tried asking why.)
Yet another example I tried, was to imagine a country where it's dishonorable, when asked about directions in a city, to say that you don't know how to get somewhere. (I remember we read and shared a laugh at such an anecdote in some book in the past.) Thus, again, you'll always get an answer, and it'll sound convincing, even if the answerer doesn't know. But again, this one didn't seem to work as good as the travel guide one; but for now I'm still keeping it to try with others in the future if needed.
PS. Ah, ok, yet another I tried was to ask them to think of the "game" of "russian roulette". You roll the barrel, you press the trigger, nothing happens. After a few lucky tries, you may get a dangerous, false feeling of safety. But then suddenly you will eventually get the full chamber.
I also tried to describe "AIs" (i.e. LLMs) as taking a shelf of books, passing them through a blender, then putting the shreds in some random order. The result may sound plausible, and even scientific (e.g. if you got medical books, or physics textbooks). The less you know the domain the books were about, the more convincing it may sound, and the harder it is to catch bullshit.
The last two pictures may have gotten some reception, but I'm not super sure, and there was still arguing especially around the books; and again, they were less of a hit than the tourist guide story.
I'm super curious if you have some analogies of your own that you're trying to use with friends and family? I'd love to steal some and see if they might work with my friends!