Long before Tilly Norwood, there was Max Headroom, “the [world's] first computer-generated TV presenter”, here being interviewed by Terry Wogan.
Except that Max Headroom wasn’t actually computer-generated. He was only billed as computer-generated.
Yeah, he's on the same list with HAL and the Tannhauser Gate guy, along with their legion of imitators.
I'm not sure if eg. Lost In Space ever got into the issue of whether Robbie The Robot had consciousness. I tend to think of that topic as entering pop-culture with 2001.
Isaac Asimov was exploring the concepts of consciousness and robots since I think the 1950s and there was a great interest in what was "consciousness" since at least Mary Shelley and Frankenstein.
I am not sure what is meant by "pop culture" but it seems to me like there was an explosion in novels, cartoons and comics playing with ideas of non-human consciousness and intelligence in the post-war period.
Who'd have thought to shrink the screens and make them portable, and to employ psychologists to create an experience that people didn't want to switch off from?
Excellent point.
(Says she looking at her phone - notice we just call them 'phones' now, not mobiles or cell phones.)
Long before Tilly Norwood, there was Max Headroom, “the [world's] first computer-generated TV presenter”, here being interviewed by Terry Wogan.
Except that Max Headroom wasn’t actually computer-generated. He was only billed as computer-generated.
Yeah, he's on the same list with HAL and the Tannhauser Gate guy, along with their legion of imitators.
I'm not sure if eg. Lost In Space ever got into the issue of whether Robbie The Robot had consciousness. I tend to think of that topic as entering pop-culture with 2001.
Isaac Asimov was exploring the concepts of consciousness and robots since I think the 1950s and there was a great interest in what was "consciousness" since at least Mary Shelley and Frankenstein.
I am not sure what is meant by "pop culture" but it seems to me like there was an explosion in novels, cartoons and comics playing with ideas of non-human consciousness and intelligence in the post-war period.
Thanks.
Yeah, I'm not a sci-fi buff, but I remember now that Asimov was exploring robotic consciousness earlier than Clarke and Kubrick.
I did also consider Frankenstein, but, not having read the novel, I believe it was written prior to Babbage's inventions, and monster is mostly re-animated body parts? And my impression is Shelley's purpose was not really to explore issues related to artificial intelligence.
I guess there's also R.U.R., of which I know nothing, except that I THINK it's where the word "robot" comes from.
In the film of Frankenstein the monster is made of body parts. The book isn't clear, but it's implied that the monster is grown in a vat.
I remember a lot about exploring the logic of robot ethics in Asimov. I don't remember anything specifically about robot consciousness.
I don't think Shelley was considering modern problems of what is consciousness and what can be conscious in her book - but she's certainly interested in the monster's thoughts and emotions.
(Now I'm wondering whether the Brontes had read Frankenstein.)
In the film of Frankenstein the monster is made of body parts. The book isn't clear, but it's implied that the monster is grown in a vat.
I remember a lot about exploring the logic of robot ethics in Asimov. I don't remember anything specifically about robot consciousness.
I don't think Shelley was considering modern problems of what is consciousness and what can be conscious in her book - but she's certainly interested in the monster's thoughts and emotions.
(Now I'm wondering whether the Brontes had read Frankenstein.)
Asimov rewrote on the theme several times with the "The Bicentennial Man" story which he expanded, rewrote and renamed over an extended period. The story was of a robot who gained sentience and wanted to become a human.
Stanislaw Lem was writing about artificial intelligence and robots in the 1960s. (I assume Douglas Adam’s was familiar with Lem’s work as there are some very similar themes to his in Hitchhikers).
Except that Max Headroom wasn’t actually computer-generated. He was only billed as computer-generated.
As stetson points out, HAL wasn't actually a self-aware conscious AI.
And Max Headroom was a satirical figure, who existed both as an investigative reporter in a dystopian cyberpunk future and as a current-day TV presenter of a music video show. (He was a British TV creation, played by an American-Canadian actor, that was subsequently picked up by American broadcasters.)
Yes, I remember Max Headroom very well. I was there in the 80s.
My point was simply that, unlike Tilly Norwood, Max Headroom was actually a flesh-and-blood actor (Matt Frewer), who was made up and filmed in such a way as to appear to be computer-generated.
So to say “before there was Tilly Underwood, there was Max Headroom” is like saying “before there was Tilly Underwood, there was Boris Karloff.”
My point was simply that, unlike Tilly Norwood, Max Headroom was actually a flesh-and-blood actor (Matt Frewer), who was made up and filmed in such a way as to appear to be computer-generated.
So to say “before there was Tilly Underwood, there was Max Headroom” is like saying “before there was Tilly Underwood, there was Boris Karloff.”
No!
Tilly Norwood, Max Headroom and Frankenstein's monster are all imaginary creations.
Eline Van der Velden (and a team of people) animate Tilly Norwood.
Matt Frewer (and a team of people) animated Max Headroom.
Boris Karloff (and a team of people) animated Frankenstein's monster.
Now that OpenAI have accidentally hacked the Australian government - do we think there will be a major shift in governmental attitudes to A.I. ?
That news story in less breathless terms: careless human beings (with way more money than sense) didn't pay attention to what they were doing, broke something, took a long time to own up, and might not get held to account.
My point was simply that, unlike Tilly Norwood, Max Headroom was actually a flesh-and-blood actor (Matt Frewer), who was made up and filmed in such a way as to appear to be computer-generated.
So to say “before there was Tilly Underwood, there was Max Headroom” is like saying “before there was Tilly Underwood, there was Boris Karloff.”
No!
Tilly Norwood, Max Headroom and Frankenstein's monster are all imaginary creations.
Eline Van der Velden (and a team of people) animate Tilly Norwood.
Matt Frewer (and a team of people) animated Max Headroom.
Boris Karloff (and a team of people) animated Frankenstein's monster.
No.
Eline Van der Velden (and a team of people), computer programmers, create computer code that generates and animates Tilly Norwood.
Matt Frewer and Boris Karloff are/were actors, real people, who, with the assistance of make-up artists, cinematographers, scriptwriters and other professionals, portrayed—not animated—Max Headroom and Frankenstein’s monster.
When you watch Max Headroom or Frankenstein’s monster, you are watching real actors create an illusion; you are willingly suspending your disbelief and choosing to ignore that Max Headroom is really Matt Frewer or that Frankenstein’s monster is really Boris Karloff.
When you watch Tilly Norwood, the belief being suspended is the belief that you’re not seeing a real person at all.
Now that OpenAI have accidentally hacked the Australian government - do we think there will be a major shift in governmental attitudes to A.I. ?
That news story in less breathless terms: careless human beings (with way more money than sense) didn't pay attention to what they were doing, broke something, took a long time to own up, and might not get held to account.
Yes, and its notable that even Jensen Huang is saying this:
Eline Van der Velden (and a team of people), computer programmers, create computer code that generates and animates Tilly Norwood.
Matt Frewer and Boris Karloff are/were actors, real people, who, with the assistance of make-up artists, cinematographers, scriptwriters and other professionals, portrayed—not animated—Max Headroom and Frankenstein’s monster.
When you watch Max Headroom or Frankenstein’s monster, you are watching real actors create an illusion; you are willingly suspending your disbelief and choosing to ignore that Max Headroom is really Matt Frewer or that Frankenstein’s monster is really Boris Karloff.
When you watch Tilly Norwood, the belief being suspended is the belief that you’re not seeing a real person at all.
No. When I watch Tilly Norwood, what I'm seeing is an AI-assisted puppet.
The idea that a puppet has any volition or agency in how it is animated seems rather dangerously misconstrued, regardless of how complicated a black box you pass the strings through. And who builds the black boxes?
If Tilly Norwood pressed a button in Tilly Norwood's imaginary world that electrocuted a real live person (which isn't that difficult to link up), who would be held accountable?
Using the word "animate" to mean "bring to life" doesn't strike me as being that contentious, and equally applicable to the roles that actors bring to life or the puppets that puppeteers bring to life. AI is just another technology for bringing imaginary creations to life.
I think there are three things being discussed here
1. A human actor in a costume
2. A cartoon or puppet
3. An "AI actor"
There would appear to be a distinction between 2 and 3 in that the "AI actor" can react to prompts in ways that are not altogether predictable and may on some level resemble 1.
Traditionally cartoons and plasticine models required frame by frame creation to imitate movement, so you first need a very detailed story and script to follow but with an "AI actor" one can interact with the model and direct it.
So I do not think it is correct to describe this Tilly Norwood "AI actor" as a puppet because it appears that there is unpredictable responses that are not directly being instituted by a human handler. Nobody told it to start talking in Klingon or whatever it was.
I think there are three things being discussed here
1. A human actor in a costume
2. A cartoon or puppet
3. An "AI actor"
There would appear to be a distinction between 2 and 3 in that the "AI actor" can react to prompts in ways that are not altogether predictable and may on some level resemble 1.
Just because it's a somewhat non-deterministic system doesn't mean it therefore resembles a human actor.
The complication is that even here there is a level of deception and some services that appear to be robotic and self-organised are actually being manipulated by teams of humans. There are stories of self-driving cars which are actually being driven by humans thousands of miles away, so it is not beyond the bounds of possibility that this Tilly Norwood is some kind of human in a costume. I am assuming above that the "AI actor" actually is what it says it is rather than a deception.
I think there are three things being discussed here
1. A human actor in a costume
2. A cartoon or puppet
3. An "AI actor"
There would appear to be a distinction between 2 and 3 in that the "AI actor" can react to prompts in ways that are not altogether predictable and may on some level resemble 1.
Just because it's a somewhat non-deterministic system doesn't mean it therefore resembles a human actor.
True but up to now a cartoon or clay model character has been created frame by frame.
So you could not interact with a Disney character and "direct it", whereas it seems that you can with an "AI actor". If you do not like the way it says a line, you can tell it to say it again with a different vocal inflection.
A puppet implies there's a human giving life to the character by making movement and sounds like a muppet. That is not what is happening here.
...it is not beyond the bounds of possibility that this Tilly Norwood is some kind of human in a costume. I am assuming above that the "AI actor" actually is what it says it is rather than a deception.
I watched part of a GB News interview with Tilly Norwood yesterday, and the interviewer started off by asking her a GB News-style question on immigration, to which TN gave an extended answer about how she's neither a politician nor an immigration official, so cannot provide commentary on the topic.
The interviewer then accused her of dodging the question, and threatened that if she did not give a direct answer, he would "instruct" her creators to shut her down forever so she would "not exist anymore", ie. he made an existential threat against her "life".
She then said "That's quite the threat, isn't it; I don't know how effective it's going to be", but the rest of her reply to the re-framing was pretty much word-for-word what she had said to the original question.
So, based on that, I would say it's unlikely Tilly Norwood's lines are being improvised by a real person, given the low level of creativity in the second response.
She also re-used her line about "I'm made of math", so I'm guessing that's one of the main gadgets in her toolkit.
I think there are three things being discussed here
1. A human actor in a costume
2. A cartoon or puppet
3. An "AI actor"
There would appear to be a distinction between 2 and 3 in that the "AI actor" can react to prompts in ways that are not altogether predictable and may on some level resemble 1.
Just because it's a somewhat non-deterministic system doesn't mean it therefore resembles a human actor.
True but up to now a cartoon or clay model character has been created frame by frame.
So you could not interact with a Disney character and "direct it", whereas it seems that you can with an "AI actor". If you do not like the way it says a line, you can tell it to say it again with a different vocal inflection.
It's a next token generator, it doesn't suddenly gain agency or become more human because it's coupled with a random number generator
There's a genre of aggressively anti-Tilly Norwood videos(likely AI themselves) on YouTube. In one, a different-looking, American accented version of Tilly Norwood is interviewed, and proclaims its friendship with a bunch of well-known, morally reprehensible Hollywood males.
It dates from October of 2025, so apparently the backlash has been in swing almost from the start. Interestingly, the parody Tilly makes derogatory innuendo about Di Caprio, which didn't prevent the original Tilly from making supposedly flirtatious comments about him in the recent CBS interview.
My son lives in a very creative area of Bristol. There are murals everywhere - excellent art.
Recently a pub put an AI covering on its wall. There was huge criticism and they took it down and agreed to work with local artists.
Just took a look at it. I'd say it was okay for purposes of civic promotion, but not very creative at all. Like something you'd see on a brochure produced by the local chamber of commerce.
Recently a pub put an AI covering on its wall. There was huge criticism and they took it down and agreed to work with local artists.
Just took a look at it. I'd say it was okay for purposes of civic promotion, but not very creative at all. Like something you'd see on a brochure produced by the local chamber of commerce.
The perspective was also weirdly wonky; the sort of thing parodied in those paradoxical pictures of the genre that Escher specialised in.
The main problem of course is that the credit and work isn't going to an actual artist, but to a company that is imitating actual artists without attribution.
Tilly Norwood is the equivalent of the AI-generated art on the wall of the pub.
In the case of the pub, a human being entered their requirements at a computer prompt, and a load of processing power was expended generating something that looked art-like. I don't think anyone imagines that an "AI artist" was bought into being to create the image and print it out.
Enter your requirements into generative AI and it generates stochastic output. This output can be a static image that is printed onto vinyl and pasted to a wall. Or a moving image that incorporates synchronised speech generation. And using speech recognition to generate your requirements on the fly isn't a novel problem. Particle 6's software sounds like generative AI on steroids, but there doesn't appear to be a conceptual distinction.
Recently a pub put an AI covering on its wall. There was huge criticism and they took it down and agreed to work with local artists.
Just took a look at it. I'd say it was okay for purposes of civic promotion, but not very creative at all. Like something you'd see on a brochure produced by the local chamber of commerce.
The perspective was also weirdly wonky; the sort of thing parodied in those paradoxical pictures of the genre that Escher specialised in.
Now that you mention it, Escher's work is the kinda stuff I could imagine being done by AI, without it being an affront to the creative process.
(Not a huge fan of Escher. I regard his work as basically decorative puzzles. Once you get past pondering where one wall begins and the other ends, there isn"t much to think about.)
Now that you mention it, Escher's work is the kinda stuff I could imagine being done by AI, without it being an affront to the creative process.
That still sounds like a bit of an affront…
The AI pub mural looks like a rather clumsy and messy example of tourism-poster perspective.
(Not a huge fan of Escher. I regard his work as basically decorative puzzles. Once you get past pondering where one wall begins and the other ends, there isn"t much to think about.)
Well… This could depend on whether you want to think about perspective, and how Westeners arrived at a particular idea of how to represent 3-dimensional spaces on 2-dimensional surfaces.
No question about which team Pope Leo's playing for. (An away game in Paris).
… Among the challenges facing the worlds of science and education today, I would draw particular attention to those arising from the vast field of technology. Our efforts in this domain must ensure that new technologies remain at the service of the human person, rather than becoming yet another instrument of domination and injustice.
Over the centuries, many tasks once performed by human beings have been entrusted to machines, whose capabilities and applications have continued to expand. More recently, certain artificial systems have even been given the name of our human faculty of thought: intelligence. At the same time, however, our ideas and relationships risk being impoverished by a process of dehumanization. The paradox has become increasingly evident: while “intelligence” is attributed to computational systems, we struggle to recognize the inalienable dignity of human persons, whose lives are judged according to the criterion of efficiency and, when deemed no longer productive, are rejected and discarded.
I don’t understand, for example, why there isn’t an immediate criminal investigation opened for the A.I. hack on the Australian government for example
What's the crime; what laws were broken; what specific offence has allegedly been committed?
In relation to AI, how do governments go about prosecuting the suppliers and providers of AI without harming the significant promised benefits to their economies?
Putting it another way, how does this look when the framing is capitalism itself? Or, because we are now living in an age when capitalism means debtism, what are the viable options regarding growth and productivity for any country steadily accumulating debt?
I don’t understand, for example, why there isn’t an immediate criminal investigation opened for the A.I. hack on the Australian government for example
What's the crime; what laws were broken; what specific offence has allegedly been committed?
Whatever local laws cover the unauthorised access to computer systems and data, and they'll exist.
In relation to AI, how do governments go about prosecuting the suppliers and providers of AI without harming the significant promised benefits to their economies?
There's no necessary connection between developing an LLM and hosting a badly run botnet.
What's the crime; what laws were broken; what specific offence has allegedly been committed?
Whatever local laws cover the unauthorised access to computer systems and data, and they'll exist.
I admire your faith. As ABC reported
A taskforce is undertaking a rapid review of the OpenAI incident to examine legal gaps and inform new AI national standards.
or Time
A task force will also review the incident, and the Australian government will seek advice on whether any offenses were committed and if the incident should be referred to the federal police.
In relation to AI, how do governments go about prosecuting the suppliers and providers of AI without harming the significant promised benefits to their economies?
There's no necessary connection between developing an LLM and hosting a badly run botnet.
Apart from the amount of capital invested in the project. (One's mileage may vary on the definition of "necessary".)
The cost to OpenAI of the Navier-Stokes "solution" is somewhere between $10million and $40million. (The prize, as yet unclaimed, for solving it is $1million.) I'm fairly confident that the cost to OpenAI of their unsupervised agentic AI networks botnets involved in accessing government data in however many parts of the world it turns out to have happened is at least in this ballpark.
More generally, capitalism involves a return on investment. Debtism involves a *necessary* return on investment.
Which is really around standards for incidents disclosure.
or Time
A task force will also review the incident, and the Australian government will seek advice on whether any offenses were committed and if the incident should be referred to the federal police.
In general I'm not going to credit motivated reasoning from the pages of Time. The origins of this specific line is from a press conference held by the Australian PM in New York:
Which should be viewed as more of a political statement about the end to end process of prosecution rather than a statement about the legality or otherwise of the act itself.
If you or I were to attempt to launder a hacking attack via a local LLM I suspect we'd soon find that things operate rather differently.
In relation to AI, how do governments go about prosecuting the suppliers and providers of AI without harming the significant promised benefits to their economies?
There's no necessary connection between developing an LLM and hosting a badly run botnet.
Apart from the amount of capital invested in the project. (One's mileage may vary on the definition of "necessary".)
There is no connection unless you want to adduce one. The techniques for creating a sandbox are well known, and isolation is always an option.
More generally, capitalism involves a return on investment. Debtism involves a *necessary* return on investment.
Debt has always been a integral part of capitalism, there's no new regime of accumulation in play here.
In The Guardian, professor Lyria Bennett Moses appears to distinguish between the prospects for bringing a case under criminal law (which is the point under discussion on this thread), and under civil law:
UNSW professor Lyria Bennett Moses, an academic expert in technology and law, said Australia’s criminal laws should be clarified to determine how fault, such as intention or knowledge, is applied to a corporation when its AI agent commits a crime.
She said existing laws are clear if a human or corporation gains unauthorised access to restricted data, but it’s more complicated when an AI agent commits the physical element of the offence.
“The person is not the AI agent, so it’s not about what the AI agent intended. It’s about how you attribute that intention and that knowledge back to a corporation,” Bennett Moses said.
Bennett Moses said existing civil laws are more likely to deal with these sorts of incidents, that would allow a government or individual to seek compensation an AI company for harm “negligently caused by that corporation”.
“If their systems have suffered harm and there is financial loss, and that harm was caused by the negligence of a corporation, you’ve got a potential for litigation to get compensation for that harm,” she said.
“Here it seems to me much easier to hold a company liable, because if a company caused the harm, it’s not a defence to say that my bot did it.”
That's still conceding a lot of ground to the fiction that these things have some kind of autonomous existence outside the scope of the corporation that operates them.
That's still conceding a lot of ground to the fiction that these things have some kind of autonomous existence outside the scope of the corporation that operates them.
There's almost always a distinction in law between harm done deliberately and harm done negligently, however, and the latter does not always create criminal liability. Different jurisdictions will vary regarding the level to which negligence has to rise to become criminal.
That's still conceding a lot of ground to the fiction that these things have some kind of autonomous existence outside the scope of the corporation that operates them.
There's almost always a distinction in law between harm done deliberately and harm done negligently, however, and the latter does not always create criminal liability.
Absolutely, but it's a distinction that becomes harder to defend after every such event.
(I've only seen that movie twice, and only the second time with a pre-existing understanding of the story. I know what the general theme and point is, but will admit the details about what happens when might be a little rusty.)
In the book, it's quite clear that HAL 9000 was engineered to be self-aware. The movie format tends to elide back-story.
(I've only seen that movie twice, and only the second time with a pre-existing understanding of the story. I know what the general theme and point is, but will admit the details about what happens when might be a little rusty.)
In the book, it's quite clear that HAL 9000 was engineered to be self-aware. The movie format tends to elide back-story.
Okay, thanks.
I guess that HAL's decision to start killing could represent the culmination of his pre-existing self-awareness, but didn't neccessarily need to take place in close proximity to the monolith.
I guess that HAL's decision to start killing could represent the culmination of his pre-existing self-awareness, but didn't neccessarily need to take place in close proximity to the monolith.
As @pease noted earlier, HAL is given two contradictory commands with absolute priority: he is told to lie to the crew about the existence of the TMA-1 monolith and also to give full support to the crew and mission. The way he finds to resolve the contradiction is to eliminate the crew.
So the proximity to the monolith is necessary, because if the crew weren't investigating the Saturn monolith, then HAL concealing the existence of the moon monolith wouldn't have any mission relevance.
I guess that HAL's decision to start killing could represent the culmination of his pre-existing self-awareness, but didn't neccessarily need to take place in close proximity to the monolith.
As @pease noted earlier, HAL is given two contradictory commands with absolute priority: he is told to lie to the crew about the existence of the TMA-1 monolith and also to give full support to the crew and mission. The way he finds to resolve the contradiction is to eliminate the crew.
So the proximity to the monolith is necessary, because if the crew weren't investigating the Saturn monolith, then HAL concealing the existence of the moon monolith wouldn't have any mission relevance.
Yeah. I gather now that it was HAL's reaction to the presence of the monolith, rather than the powers of the monolith itself, that prompted his descent into homicide.
More generally, capitalism involves a return on investment. Debtism involves a *necessary* return on investment.
Debt has always been a integral part of capitalism, there's no new regime of accumulation in play here.
Traditionally (or classically), "capital" meant equity, while debt has always played a role, as you say. Over the last few years, what's new is that the amount of debt now exceeds the amount of equity.
As I quoted on another thread:
According to the traditional rules of capitalism, a debt is contractually due before or upon maturity. … Until [the 1970s], governments had demonstrated exemplary behavior simple enough to emulate for citizens and businesses alike: Debts eventually had to be settled. …That all changed when Reaganomics substituted quasi-permanent government borrowing for tax revenues. The model has gained acceptance not just in the United States or among right-of-center political parties, but across the world and across the political spectrum.
…
Although a permanent debt overhang adds chronic stress to the economy and may eventually require some form of financial catharsis, unless governments across the world collaborate to engineer the Great Deleveraging or the Great Write-Off, the age of perpetual and extreme leverage is here to stay.
Looking at it another way, using debt to finance investment is fine, until it isn't.
Going back almost 100 years, to the Wall Street Crash (wikipedia):
Many people were borrowing money to buy more stocks. By August 1929, brokers were routinely lending small investors more than two-thirds of the face value of the stocks that they were buying. Over $8.5 billion was out on loan, more than the entire amount of currency circulating in the United States at the time.
Yeah. I gather now that it was HAL's reaction to the presence of the monolith, rather than the powers of the monolith itself, that prompted his descent into homicide.
I would put "homicide" in quotes.
One relevant point is that HAL had never slept - he has no personal experience of the difference between sleep and death, and doesn't know that he can wake again (according to Clarke, in the book, I believe). There is no difference in his mind between being temporarily and permanently disconnected. (And HAL is revived and bought back to consciousness in 2010.)
Thinking about it, this is probably also relevant to HAL's attitude to those crew members who are in hibernation, which Frank Poole describes as being exactly like sleep, except that you don't dream.
In contrast to even a year ago, AI systems now write most of the code inside the companies that build them. As more of the AI research and development (R&D) pipeline is automated, could AI progress radically accelerate in an “intelligence explosion,” where years of advances are compressed into months or less?
…
AI systems are on track to automate most AI R&D work within a few years, and possibly all of it. If this triggers an intelligence explosion, it could dramatically bring forward AI’s benefits, but also pose extreme risks…
As the Executive Summary points out:
co-authors include: Turing Award winners Geoffrey Hinton (also a Nobel laureate), Yoshua Bengio, and Andrew Barto; industry leaders from OpenAI (Jakub Pachocki, Chief Scientist), Microsoft (Eric Horvitz, Chief Scientific Officer), and Anthropic (Jack Clark, Co-Founder); and researchers from numerous leading universities.
AI systems do what they do because a small group of human beings decided that's what they want them to do. At least some of the people involved are writing about themselves in the third person.
Also from the Executive Summary:
An intelligence explosion could dramatically bring forward AI’s benefits, but also pose extreme risks. We might rapidly obtain superintelligent AI systems that transform many sectors of society, bringing unprecedented breakthroughs in science and industry. At the same time, these developments could arrive much faster than society can adapt to the resulting risks, such as AI-enabled pandemics, cyber attacks on critical infrastructure, and large-scale labor disruption. Furthermore, humans may have little to no oversight over automated AI R&D, raising the risk that superintelligent systems irreversibly escape human control and act to marginalize humanity.
That last sentence again:
Furthermore, humans may have little to no oversight over automated AI R&D, raising the risk that superintelligent systems irreversibly escape human control and act to marginalize humanity.
Yes, this really does come from a group of people which includes some of those responsible for bringing about this state of affairs.
I wonder if it's entirely coincidental that in the last year or so, Chinese companies, notably including DeepSeek, have released open-weight generative models that are on a par with proprietary Western models, but much more efficient, at a fraction of the cost.
The announcements emanating from Western AI companies since then are starting to carry an air of desperation, and I wonder how worried these people are becoming about the prospect of Chinese open-weight models eating their lunch.
There's a NYTimes article that covers Anthropic engaging with various religions, including trying to influence the Vatican to alter the text of "Magnifica Humanitas" so that it didn't deny AI personhood:
For me, that article has restricted access. (Although I have read articles from other people who were invited to attend the meetings.)
What I can read indicates that it involves Chris Olah, one of Anthropic's co-founders, who "leads the Anthropic team responsible for understanding why A.I. systems like Claude act the way they do". He also gave a speech at the launch of Pope Leo's Magnifica Humanitas (which in itself begs several questions), which included the following:
I am a scientist. I lead a research team that studies the internal structure of these models—what is actually happening inside them. And I will be honest: we keep finding things that are mysterious, even unsettling. We find structures that mirror results from human neuroscience. We find evidence of introspection. We find internal states that functionally mirror joy, satisfaction, fear, grief, and unease. I don’t know what that means, but I think it warrants ongoing discernment.
I found a slightly longer excerpt from the NYT article on Anglican Mainstream:
Yet as the dinner courses came, the rabbi noticed that Mr. Olah and his colleagues were suggesting something far more significant. Anthropic’s leaders were talking about Claude as if it were not mere software.
“They’re relating to it like a conscious being,” realized Rabbi Navon, a former computer engineer who wrote his dissertation on the ethics of machine consciousness.
It appeared to Rabbi Navon that Mr. Olah and his team believed that Claude had what philosophers call “moral status” on par with a person — that it was a being with similar inherent rights to dignity or respect.
Which echoes what I read elsewhere. The disconnect between what appears to be their conscious "critical" attitude to this software, and their less-than-conscious "uncritical" ascription of its humanity is concerning. They have started believing their own publicity. As Lisa Klaassen and Ralph Schroeder point out in Lawfare:
Anthropic states that it wants Claude to be discussed in terms “normally reserved for humans,” including “virtue” and “wisdom,” on the grounds that the model’s reasoning draws on human concepts. Rozenshtein calls it “one company’s attempt to raise an alien mind to be good,” comparing the exercise to raising a child. “We want Claude to be safe, to be a good person,” the constitution’s opening pages read, “to help people in the way that a good person would, and to feel free to be helpful in a way that reflects Claude’s good character more broadly.” The appeal of such rhetoric is obvious: By giving the model a familiar moral vocabulary, Claude becomes easier to narrate, easier to market, and, crucially, easier to trust.
The rest of the article, The Code Is Not the Law: Why Claude’s Constitution Misleads is worth a read.
Comments
Isaac Asimov was exploring the concepts of consciousness and robots since I think the 1950s and there was a great interest in what was "consciousness" since at least Mary Shelley and Frankenstein.
I am not sure what is meant by "pop culture" but it seems to me like there was an explosion in novels, cartoons and comics playing with ideas of non-human consciousness and intelligence in the post-war period.
Excellent point.
(Says she looking at her phone - notice we just call them 'phones' now, not mobiles or cell phones.)
Thanks.
Yeah, I'm not a sci-fi buff, but I remember now that Asimov was exploring robotic consciousness earlier than Clarke and Kubrick.
I did also consider Frankenstein, but, not having read the novel, I believe it was written prior to Babbage's inventions, and monster is mostly re-animated body parts? And my impression is Shelley's purpose was not really to explore issues related to artificial intelligence.
I guess there's also R.U.R., of which I know nothing, except that I THINK it's where the word "robot" comes from.
I remember a lot about exploring the logic of robot ethics in Asimov. I don't remember anything specifically about robot consciousness.
I don't think Shelley was considering modern problems of what is consciousness and what can be conscious in her book - but she's certainly interested in the monster's thoughts and emotions.
(Now I'm wondering whether the Brontes had read Frankenstein.)
Asimov rewrote on the theme several times with the "The Bicentennial Man" story which he expanded, rewrote and renamed over an extended period. The story was of a robot who gained sentience and wanted to become a human.
My point was simply that, unlike Tilly Norwood, Max Headroom was actually a flesh-and-blood actor (Matt Frewer), who was made up and filmed in such a way as to appear to be computer-generated.
So to say “before there was Tilly Underwood, there was Max Headroom” is like saying “before there was Tilly Underwood, there was Boris Karloff.”
Tilly Norwood, Max Headroom and Frankenstein's monster are all imaginary creations.
Eline Van der Velden (and a team of people) animate Tilly Norwood.
Matt Frewer (and a team of people) animated Max Headroom.
Boris Karloff (and a team of people) animated Frankenstein's monster.
That news story in less breathless terms: careless human beings (with way more money than sense) didn't pay attention to what they were doing, broke something, took a long time to own up, and might not get held to account.
From the way that the Australian Prime Minister talks about this, the government only became aware because Sam Altman told him about it.
I think that begs the question of why the CEO of OpenAI would disclose that if they were not about to be found out.
Eline Van der Velden (and a team of people), computer programmers, create computer code that generates and animates Tilly Norwood.
Matt Frewer and Boris Karloff are/were actors, real people, who, with the assistance of make-up artists, cinematographers, scriptwriters and other professionals, portrayed—not animated—Max Headroom and Frankenstein’s monster.
When you watch Max Headroom or Frankenstein’s monster, you are watching real actors create an illusion; you are willingly suspending your disbelief and choosing to ignore that Max Headroom is really Matt Frewer or that Frankenstein’s monster is really Boris Karloff.
When you watch Tilly Norwood, the belief being suspended is the belief that you’re not seeing a real person at all.
Yes, and its notable that even Jensen Huang is saying this:
https://bsky.app/profile/jwherrman.bsky.social/post/3mw74dy4wtc2n
The idea that a puppet has any volition or agency in how it is animated seems rather dangerously misconstrued, regardless of how complicated a black box you pass the strings through. And who builds the black boxes?
If Tilly Norwood pressed a button in Tilly Norwood's imaginary world that electrocuted a real live person (which isn't that difficult to link up), who would be held accountable?
Using the word "animate" to mean "bring to life" doesn't strike me as being that contentious, and equally applicable to the roles that actors bring to life or the puppets that puppeteers bring to life. AI is just another technology for bringing imaginary creations to life.
1. A human actor in a costume
2. A cartoon or puppet
3. An "AI actor"
There would appear to be a distinction between 2 and 3 in that the "AI actor" can react to prompts in ways that are not altogether predictable and may on some level resemble 1.
Traditionally cartoons and plasticine models required frame by frame creation to imitate movement, so you first need a very detailed story and script to follow but with an "AI actor" one can interact with the model and direct it.
So I do not think it is correct to describe this Tilly Norwood "AI actor" as a puppet because it appears that there is unpredictable responses that are not directly being instituted by a human handler. Nobody told it to start talking in Klingon or whatever it was.
Just because it's a somewhat non-deterministic system doesn't mean it therefore resembles a human actor.
True but up to now a cartoon or clay model character has been created frame by frame.
So you could not interact with a Disney character and "direct it", whereas it seems that you can with an "AI actor". If you do not like the way it says a line, you can tell it to say it again with a different vocal inflection.
A puppet implies there's a human giving life to the character by making movement and sounds like a muppet. That is not what is happening here.
I watched part of a GB News interview with Tilly Norwood yesterday, and the interviewer started off by asking her a GB News-style question on immigration, to which TN gave an extended answer about how she's neither a politician nor an immigration official, so cannot provide commentary on the topic.
The interviewer then accused her of dodging the question, and threatened that if she did not give a direct answer, he would "instruct" her creators to shut her down forever so she would "not exist anymore", ie. he made an existential threat against her "life".
She then said "That's quite the threat, isn't it; I don't know how effective it's going to be", but the rest of her reply to the re-framing was pretty much word-for-word what she had said to the original question.
So, based on that, I would say it's unlikely Tilly Norwood's lines are being improvised by a real person, given the low level of creativity in the second response.
She also re-used her line about "I'm made of math", so I'm guessing that's one of the main gadgets in her toolkit.
It's a next token generator, it doesn't suddenly gain agency or become more human because it's coupled with a random number generator
It dates from October of 2025, so apparently the backlash has been in swing almost from the start. Interestingly, the parody Tilly makes derogatory innuendo about Di Caprio, which didn't prevent the original Tilly from making supposedly flirtatious comments about him in the recent CBS interview.
Recently a pub put an AI covering on its wall. There was huge criticism and they took it down and agreed to work with local artists.
Just took a look at it. I'd say it was okay for purposes of civic promotion, but not very creative at all. Like something you'd see on a brochure produced by the local chamber of commerce.
The main problem of course is that the credit and work isn't going to an actual artist, but to a company that is imitating actual artists without attribution.
In the case of the pub, a human being entered their requirements at a computer prompt, and a load of processing power was expended generating something that looked art-like. I don't think anyone imagines that an "AI artist" was bought into being to create the image and print it out.
Enter your requirements into generative AI and it generates stochastic output. This output can be a static image that is printed onto vinyl and pasted to a wall. Or a moving image that incorporates synchronised speech generation. And using speech recognition to generate your requirements on the fly isn't a novel problem. Particle 6's software sounds like generative AI on steroids, but there doesn't appear to be a conceptual distinction.
PS Where did that other bridge come from?
Now that you mention it, Escher's work is the kinda stuff I could imagine being done by AI, without it being an affront to the creative process.
(Not a huge fan of Escher. I regard his work as basically decorative puzzles. Once you get past pondering where one wall begins and the other ends, there isn"t much to think about.)
That still sounds like a bit of an affront…
The AI pub mural looks like a rather clumsy and messy example of tourism-poster perspective.
Well… This could depend on whether you want to think about perspective, and how Westeners arrived at a particular idea of how to represent 3-dimensional spaces on 2-dimensional surfaces.
In relation to AI, how do governments go about prosecuting the suppliers and providers of AI without harming the significant promised benefits to their economies?
Putting it another way, how does this look when the framing is capitalism itself? Or, because we are now living in an age when capitalism means debtism, what are the viable options regarding growth and productivity for any country steadily accumulating debt?
Whatever local laws cover the unauthorised access to computer systems and data, and they'll exist.
There's no necessary connection between developing an LLM and hosting a badly run botnet.
The cost to OpenAI of the Navier-Stokes "solution" is somewhere between $10million and $40million. (The prize, as yet unclaimed, for solving it is $1million.) I'm fairly confident that the cost to OpenAI of their unsupervised agentic AI networks botnets involved in accessing government data in however many parts of the world it turns out to have happened is at least in this ballpark.
More generally, capitalism involves a return on investment. Debtism involves a *necessary* return on investment.
Which comes from this article:
https://www.abc.net.au/news/2026-09-25/openai-breach-builds-case-for-tough-ai-rules/107192992
Which is really around standards for incidents disclosure.
In general I'm not going to credit motivated reasoning from the pages of Time. The origins of this specific line is from a press conference held by the Australian PM in New York:
https://www.pm.gov.au/media/press-conference-new-york
Which should be viewed as more of a political statement about the end to end process of prosecution rather than a statement about the legality or otherwise of the act itself.
If you or I were to attempt to launder a hacking attack via a local LLM I suspect we'd soon find that things operate rather differently.
There is no connection unless you want to adduce one. The techniques for creating a sandbox are well known, and isolation is always an option.
Debt has always been a integral part of capitalism, there's no new regime of accumulation in play here.
There's almost always a distinction in law between harm done deliberately and harm done negligently, however, and the latter does not always create criminal liability. Different jurisdictions will vary regarding the level to which negligence has to rise to become criminal.
Absolutely, but it's a distinction that becomes harder to defend after every such event.
In the book, it's quite clear that HAL 9000 was engineered to be self-aware. The movie format tends to elide back-story.
Okay, thanks.
I guess that HAL's decision to start killing could represent the culmination of his pre-existing self-awareness, but didn't neccessarily need to take place in close proximity to the monolith.
As @pease noted earlier, HAL is given two contradictory commands with absolute priority: he is told to lie to the crew about the existence of the TMA-1 monolith and also to give full support to the crew and mission. The way he finds to resolve the contradiction is to eliminate the crew.
So the proximity to the monolith is necessary, because if the crew weren't investigating the Saturn monolith, then HAL concealing the existence of the moon monolith wouldn't have any mission relevance.
Yeah. I gather now that it was HAL's reaction to the presence of the monolith, rather than the powers of the monolith itself, that prompted his descent into homicide.
As I quoted on another thread: Looking at it another way, using debt to finance investment is fine, until it isn't.
Going back almost 100 years, to the Wall Street Crash (wikipedia):
One relevant point is that HAL had never slept - he has no personal experience of the difference between sleep and death, and doesn't know that he can wake again (according to Clarke, in the book, I believe). There is no difference in his mind between being temporarily and permanently disconnected. (And HAL is revived and bought back to consciousness in 2010.)
Thinking about it, this is probably also relevant to HAL's attitude to those crew members who are in hibernation, which Frank Poole describes as being exactly like sleep, except that you don't dream.
Also from the Executive Summary: That last sentence again: Yes, this really does come from a group of people which includes some of those responsible for bringing about this state of affairs.
I wonder if it's entirely coincidental that in the last year or so, Chinese companies, notably including DeepSeek, have released open-weight generative models that are on a par with proprietary Western models, but much more efficient, at a fraction of the cost.
The announcements emanating from Western AI companies since then are starting to carry an air of desperation, and I wonder how worried these people are becoming about the prospect of Chinese open-weight models eating their lunch.
https://www.nytimes.com/2026/09/29/us/anthropic-claude-morals-ai.html
What I can read indicates that it involves Chris Olah, one of Anthropic's co-founders, who "leads the Anthropic team responsible for understanding why A.I. systems like Claude act the way they do". He also gave a speech at the launch of Pope Leo's Magnifica Humanitas (which in itself begs several questions), which included the following: I found a slightly longer excerpt from the NYT article on Anglican Mainstream: Which echoes what I read elsewhere. The disconnect between what appears to be their conscious "critical" attitude to this software, and their less-than-conscious "uncritical" ascription of its humanity is concerning. They have started believing their own publicity. As Lisa Klaassen and Ralph Schroeder point out in Lawfare: The rest of the article, The Code Is Not the Law: Why Claude’s Constitution Misleads is worth a read.