Magnifica Humanitas

The point de départ for today's offering is this selection of quotes from Pope Leo XIV's recent encyclical:

• “A more moral AI is not enough if that morality is determined by a few. What is needed is a more active political involvement that is capable of slowing things down when everything is accelerating, and of protecting the opportunities for communities still to be able to participate and ask questions.” 
  
  

• “Indifference to the truth leads, slowly but surely, to a descent into totalitarianism. As the philosopher Hannah Arendt wrote, the ideal subjects of such regimes are not so much  who are ideologically convinced, but rather “people for whom the distinction between fact and fiction (i.e., the reality of experience) and the distinction between true and false (i.e., the standards of thought) no longer exist.”
  
  

• “Today, more than ever, without prejudice to the right to self-defense in the strictest sense, it is important to reaffirm that the “just war” theory, which has all too often been used to justify any kind of war, is now outdated. Humanity possesses far more effective and capable tools for promoting human life and resolving conflicts, such as dialogue, diplomacy and forgiveness.”
  
  

• “Disarming AI means freeing it from the mentality of “armed” competition, which today is not limited simply to the military context, but is also an economic and cognitive phenomenon…
  

  <AC-elision>
  This elision is made in the original selection. But I think the original reference to the mega-rich entrepre-nerds is worth repeating.
  

  ...phenomenon. This entails a race for ever more powerful algorithms and larger datasets, driven by the desire to secure geopolitical or commercial dominance. To disarm means discrediting the assumption that technical power automatically confers the right to govern. To disarm does not mean rejecting technology, but preventing it from dominating humanity. It means freeing technology from monopolistic control and opening it to discussion and debate, therefore making it human-friendly and restoring it to the plurality of human cultures and ways of life. Our task today...

  He's a good man, our Leo. 

  </AC-elision>

  ...Our task today is not only ethical or technical. It is ecological in the deepest sense, for it concerns a new dimension of our common home. AI is already an environment in which we are immersed, as well as a force with which we must engage. For this reason, merely regulating it is insufficient; it must be disarmed, welcoming and accessible.”
  
  

• “So-called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships and do not know from within what love, work, friendship or responsibility mean. Nor do they have a moral conscience, since they do not judge good and evil, grasp the ultimate meaning of situations, or bear responsibility for consequences. They may imitate language, behavior and analytical skills, or even simulate empathy and understanding, but they do not understand what they produce, for they lack the affective, relational and spiritual perspective through which human beings grow in wisdom.”

The selection was made by Alastair Campbell, who famously 'doesn't do God', which accounts for His (sic) absence from the chosen text. The encyclical itself (which Campbell recommended in a recent-ish The Rest Is Politics) contains 97 references to 'God' (not counting a further 2 in the contents). So this selection is quite impressive: I would be hard-pressed to choose so many God-free words from such a God-rich document

L'Envoi 

I know that AI and spellcheckers aren't the same thing, but giving the computer more initiative than it deserves is a temptation common to both. And it's at least possible that an AI agent generated this bit of spam and called on a spellchecker in the process:

You won't be surprised to learn that the mail with this header went straight to the bin, so I don't know what largesse I've missed out on, but as with John Collier – who I remember accusing (in my long distant clothes-buying days) of 'catering only for fat dwarves' – I don't think I've missed much

There's more to come on this subject, but that's enough for now.

b

The Sapiens and the cow

In a fairly recent ...

<tangent>
OK, 17 September. There comes a time in the mid-seventies when recentness becomes a somewhat flexible measure

</tangent>

...(members only – so here's a link that will no doubt invite disbursement of some kind, if it works at all) extra edition of The Rest Is Politics Plus, which has a conversation with Rory Stewart  about (you guessed it) AI (the man's obsessed), Yuval Noah Harari discusses a consequence of the break-neck unregulated development of AI. In a clear-headed and unsensational vision of the future, ignoring for the time being the possibility of malign  intervention and mass extermination, he says that – in a world where the entrepre-nerds have a high enough Inanity Quotient  to let AI achieve Recursive Self Improvement (AIs making the next generation of AI's ever further from human understanding or control) – 'We'll be like the other animals on the planet'. He imagines a conversation with a cow.

<parable verbatim="no, but I think I got the gist", embellishment="What do you think?"> 

'You [the cow] are part of a world economic system involving markets in stocks and bonds operated by algorithms that are far beyond your knowledge or control.' 

The cow says 'What are stocks? What are bonds? I know what grass is; I know what fences are; I know that man looks after me and exercises some kind of mastery over me and my kind.'

Comes a time when market conditions change and the supra-national conglomerate that owns the cow goes bust. You tell the cow 'This is the end of the road.' 

'Aha,' thinks the cow, 'I know what a road is – those hard bits of land that are uncomfortable to wal...'.  And the abbattoir's van pulls up.

</parable>

Just over a year ago, The Pimlico Journal published a thought-provoking article:

<j'accuse>

MPs are almost certainly [HD: that is, absolutely certainly – look at the figure It's just hard to prove,] using ChatGPT to generate Commons speeches

A statistical analysis, and some tell-tale signs

Pimlico Journal

Sep 01, 2025

...

I. The statistical evidence for believing that MPs are regularly using ChatGPT to generate Commons speeches

...

What would be far more egregious and alarming [HD: than the use of ChatGPT for constuency work, as defended here by Mike Reader, MP] is the use of AI to write parliamentary speeches. Zoe Crowther of PoliticsHome posted a chart showing a recent spike in the frequency of the phrase ‘I rise to speak’ in Hansard after being tipped off by Reader  that this is a tell-tale sign of ChatGPT.

This spurred further debate, with The Times publishing an editorial on Friday with a similar view to my own: sometimes acceptable if responding to emails, not so much if being used in Parliament. The editorial ended with the question: ‘What cruel news channel will it be that splices together examples of this synthetic oratory?’

Pimlico Journal is not a news channel, but we hope to be able to rise to this challenge nonetheless. In order to do so, we will start by looking for further statistical indications that AI is being used to generate speeches in Parliament. After all, ‘I rise to speak’ is just one phrase, and with a large crop of fresh MPs, it is not outside the realms of possibility that the phrase just caught on through a sort of social contagion because newcomers thought it was somehow ‘the done thing’.

...

The following search terms were sourced by browsing articles and online discussions on what words and phrases AI-generated text overuses, as well as my own experience (some of these words, despite often being cited as disproportionately frequent in AI-generated text, have been excluded for reasons that are listed in ...

<HTTPS: 7, HD:0>
Sorry. I tried to put a link here, but the Substack site involved me in a Kafka-esque verification loop that even I grew tired of.
</HTTPS: 7, HD:0>

...Appendix A). Below, you’ll find a series of graphs showing ... the frequency of AI-associated words and phrases, as well as some reference words, in the House of Commons each year from 2007 to 2025. The frequencies of the word or phrase in each year were adjusted by a factor (Appendix B)...

<HTTPS: 7, HD:0 ALT="Ditto" />

... that attempted to measure how many words were spoken in total in Parliament in that year: for instance, 2025 is adjusted upwards because there have been fewer words spoken in the eight months of the year so far compared to most full years.

<glossary>
The Z-score is a measure of how far a value in a set is from the mean, expressed in units of the standard deviation. And if you know what that means...

<glossary>

</j'accuse>

So this thing is here and is already worming its way into our culture. And I can't help thinking that homo sapiens is the cow in Yuval Noah Harari's parable, driven to extinction by a super-inelligent AI with RSI, just as all the other homo species were driven to extinction by the (possibly well-intemtioned) spread of homo sapiens.

This illustration from The Language Puzzle sprang to mind.
I'm afraid my screen-grab cut off

Ardipithecus ramidus's chin

[no, this isn't Ardipithecus ramidus agnathus]

L'Envoi

That illustration came to mind as a result of an experiment I've been doing for the past year, putting Audible through its paces; in particular I wanted to see how useful Audible was as 'background music' to reading (my eyesight being less than perfect).

I wrote a review, 'published' within Audible...

<air-quotes>
You may see from my review that the Audible app seems ro me to be lashed together with chewing gum and baler twine, so published in this case seems not to mean 'made available as a publication'.
</air-quotes>

Here it is; blest if I can find it in Audible.

<review justification="I SHALL get it read somewhere>

I wish Audible would wait until their readers could read the actual printed book as published; not doing this is a disservice to the authors and also to the listeners  - who get a distorted view of what the author actually meant. I first noticed this in The Book of Why, which I listened to while scanning the printed text. There were numerous cases where the reader changed the written words (usually in ways that had little or no effect on the meaning - for example,  'and so on' for 'etc', 'which' for 'that'). More serious, as the reader had a final page proof, often a piece of text - when the author had added or removed, or reworded something - was omitted or added or unreworded.

In the case of The Language Puzzle this was more obvious to me as I studied Linguistics at degree and post-graduate level, with the added awareness that I knew how linguists speak about things - so that I was (often painfully!) aware of little infelicities. For example, the reader, when the Frenchman was first mentioned, treated the name Brocas to a lavishly Gallic pronunciation; later, presumably after clarification from the author, he used the more Anglophone rendering used by linguists. There were lots of other small points, which though trivial didn't inspire confidence; for example,his stress on homo habilis  varied - starting with homo habilis (the normal stress used by archaeologists, such as Prof. Alice Roberts - quoted on the book's jacket) but settling down on homo habilis (which I'd never heard before)

But the main problem was that at the page-proof stage there had been a full-scale reorganization and the 16-chapter book became a 19-chapter recording, with the three extra chapters cobbled together out of bits of formerly longer chapters. At the beginning there is a breakdown of the argument to come: ' Chapter n deals with <thing>, Chapter n+1 deals with <other-thing>... etc.', and this is a bitty mess (quite hard to follow, as it has to make reference to all the disparate cobbled-together bits). But in the final chapter there is a similar resumé which  is suspiciously unbitty. I didn't check the references, but it feels as though the reorganization escaped the editor's notice. In short, the table of contents is a work of fiction.

The reader had a sing-song voice - not monotonous, but perhaps oligotonous (involving a small number of tones - which he assigned more or less at random throughout each sentence).

My score for 'story' refers to the argument - which I found unconvincing. Of one writer he wrote 'this seems to me a speculation too far' - a criticism that I often felt applied to his own surmises. For example, 'a biface hand-axe could be carried around and used when needed,, just like words' (not verbatim, but that was what he was suggesting); or a chimp's grunts show the beginnings of syntax - which is true, but only in a metaphorical sense. And his frequent resort to synaesthesia as an 'explanation' struck me as fanciful (bordering on risible).

My down-marking for 'Story' was only slightly mitigated by the fact that the sources used were so eclectic and interesting (although the conclusions drawn were so extravagant); I say 'only slightly' because the Audible version doesn't include the footnotes (or perhaps does include them somewhere not pointed to by the lamentable table of contents.

Listen to this book if you want glimpses of an impressive range of archaeological and ethnological and biological and genetic …(etc) findings, though I'm not sure this book adds much to an understanding of the development of natural language.

</review>

Enough!

b