Computer says ON

Last December, before Elon Musk started wielding his billion dollar chainsaw, an article in Futurism...              

<parenthesis subject="The link not followed">
( I had, incidentally, published a link to it in an article in The Byte quoted in this:

AI models routinely hallucinate and make up facts. They have no understanding of language, but instead use statistical predictions to generate cogent-sounding text based on the human writing they've ingested 

But I'd made the mistake of leaving a link in that I hadn't followed. Usually I remove any links that I've not followed; it seems to me to be a requirement of responsible blogsmansanship – making sure my blog doesn't lead people astray.) 
</parenthesis>

... pointed out that information seen through the distorting lens of Elon Musk's imagination (EMformation?) might not be entirely reliable: 

The astonishingly gullible billionaire ...

<parenthesis type="silly quip">
(a sequel to The Very Hungry Caterpillar?) 
</parenthesis>

...has a well-documented tendency to spread misinformation, an embarrassing and often dangerous quality that's undermined his standing in the world of science and technology.

In fact, even his own AI company's chatbot Grok — which Musk himself has branded as a brash truth-teller — has a pretty good grasp on the situation.

"Yes, there is substantial evidence and analysis suggesting that Elon Musk has spread misinformation on various topics, including elections, to a very large audience through his social media platform, X," Grok responded when asked if "Elon Musk spread misinformation to billions of people."

Source

 The article goes on:

Grok pointed to the wealth of misinformation Musk has shared in recent years, especially when it comes to the most recent presidential election.

"Musk's posts related to elections, which have contained misleading or false claims, have amassed billions of views," Grok wrote. "Musk has shared manipulated videos and debunked claims about voting processes, including allegations about non-citizen voting, which are common themes in misinformation narratives."

Case in point, less than a day after Musk cosigned a meme that called people who "still believe everything shown in [sic] news" dumb, he shared a video that purportedly showed "armed communist Maduro gangs... storming polling stations in Punta Cardón," Venezuela.

As many users quickly pointed out, the video actually showed thieves attempting to steal air conditioners.

This reminded me...

<parenthesis>
Not sure why. Perhaps it was that the bad uses of AI brought to mind the possible good ones. That's me – always looking on the bright side.
</parenthesis>

... of a recent addition to the The Rest Is Politics: Leading series. To quote that website's description:

How many people are going to die as a result of American support for USAID programs being removed? What are the limits of thinking of AI as the silver bullet for healthcare? How do we bring the doubling of the human lifespan to everyone alive without bankrupting our societies? 

Atul Gawande, late of USAID, and soon to be reinstated as a Professor at Harvard...

<trump-warning>
(that is, if His Royal Mendaciousness...

<inline-ps>
I chose this word to resonate with "Highness", but "mendacity" resonates almost as well with "Majesty" and is much more common, as this Ngram shows:

I wonder what happened in mid-19th century
to  make 'mendacity' so popular. So many
questions, so little time...

</inline-ps>

...Mendacity doesn't have it closed down for the bare-faced effrontery of having been founded before the USA was [and shame on all the other institutions and people [particularly politicians] kowtowing to the infantile/amoral vandal...

<tangent>
[Meanwhile, in Ukraine, where traditional diplomats want 'a just peace' he wants 'just peace' so that he can get on with general exploitation]
</tangent>

...]). As Anthony Scaramucci says in another Goalhanger podcast, we can look forward to "a Golden Age of Corruption'.
 </trump-warning>

... is well worth a listen He talks with wit and wisdom (and, appropriately enough for a surgeon, incision) about not only Trump and his lying hench-nerd (to call Musk a henchman would be an insult to henchmen), but about the uses of AI in medicine. At one point he says that in the average clinic every hour of contact with patients requires two hours' work on documentation. 

Which adds weight to a Medscape survey on the use of AI in medicine, published last September, here

   

Of all the areas considered by respondents (745 UK based doctors), 'Administration and recordkeeping' us seen by doctors as potentially the most promising area  for using AI.

But this doesn't mean much; they might be wrong in that belief; and UK-based doctors may not be the best judges of this anyway; I don't know. Besides, AI may just be the deus ex machina that – in this pipe dream – gets rid of all the pen-pushing.

 Towards the end of the survey report we get to the main event:

That last comment is really GOSH-worthy. And don't you agree that the penultimate word is doing a lot of heavy-lifting?

But, although I can't say I'm  impressed by the survey, the message seems to be that AI can have a big influence on medicine as long as expectations are kept in check. The tech-bros waving their iPhones at Atul Gawande and saying 'This is gonna replace you' are wrong. As he says 'How is it going to deliver [a] baby?'

There's good news: 

 

<tangent>
Oh dear. That "61% vs 47%" is a symptom of the confusion and haste that characterizes this report. They should've used AI – and if they did they should have done a decent job of post-editing.
<tangent>

 Enough. But before I go I must mention the next WCS concert, at All Saints in ... oo-er. less than a month.

b

Update 2025.05.27.13:05 – Added <inline-ps />

For The Record

Earlier this week I received the latest copy ('latest' being the 2023-24 edition, celebrating 40 years of 'women at Corpus') of my college's annual  magazine, The Record. I did a quick sum and realized that this meant '40 years of women resident at Corpus'.  

<autobiographical-note>

The reason for my temporary confusion was that when I first dined in Corpus, having arrived in late September 1971 (just before my first term – the Oral exam for foreign languages being held before we'd been exposed to anything like tuition) women were admitted to dine at High Table. (This barefaced assault on the patriarchy was greeted by what Nigel Starmer-Smith might have called 'some ill-mannered hissing'). To mark the occasion there were four courses, and as this was my first experience of Formal Hall, my expectations of my second Hall were rudely dashed.  

I wrote about this (with a paranoid self-justificatory note about the timing of Cambridge terms) here.
</autobiographical-note>

Having skimmed through The Record I was about to put it out of my mind ...

<tangent>
(the role of recycling in this process is a matter for ongoing debate with MrsK)
</tangent>

...when I took a last look at the Old Members section, which passes on bits of information about alumni/-ae, listed by year of matriculation and found that I had nearly missed this:

Excerpt from 'Old Members' section of The Record

The text was vaguely familiar, but in this wordy (linkless) form almost entirely useless. Guiltily I remembered a note I had written two or three years ago, with links to posts with 'Corpus-related content'. I had looked in the 2022-23 edition to see whether my attempt at publicity had borne fruit, and then forgotten it.

But in a blog with well over 500 entries, posted over a dozen years, it would be ridiculous to draw from that entry in The Record the conclusion that the content was always related to Corpus. My subjects are many and various, often relating to language and music. I append an example  by way of a PS (or makeweight) that can safely be ignored by the non Corpuscular.

Anyway I hope Blogger's servers can handle the surge in traffic now that I've included these links:

• Chris Andrew's book (not books)
• Augustine Gospels
• Singing in Chapel
• Marlowe portrait
• Chronopage also here and here
• Wilkins (New Court architect)
• DNA Origami/Rutherford

b

PS

On BBC Radio 4 the other  day...

<correction>
This was in fact more than a year ago. I just left it unfinished and unpublished.
</correction>

 ...in a programme I didn't have time to digest, but will revisit..

<parenthesis>
(I did, but found I had no more to add)
</parenthesis>

...a Spanish-speaking vulcanologist...

 <parenthesis>
(I think they used the Globish volcanologist, but I'm a fan of Latinate derivations...

<glossary>
[volcanoes got their name from Vulcan, and those ignorant Romans didn't know squat about seismology and stuff ]
</glossary>

...at the possible expense of international understanding.)
 </parenthesis>

... talking about an eruption he had seen. And he said 'I couldn't stop to watch [the eruption]'.

His English was good enough for him to use "couldn't", but not good enough to stretch to 'couldn't stop watching' (let alone the idiomatic 'couldn't take my eyes off it').

This reminded me of a picture shared in an ESOL teachers' group on Facebook...

<autobiographical-note>
My biography in The Record doesn't mention the fact that, after the skinflints at HP laid me off a few months short of my 20th year (when they'd have had to fork out for an award), I became a language tacher.
</autobiographical-note>

...discussed here

<pre_script>

[This] recalled for me a diagram (I won't say picture) that I used to use to show the difference between "stop + infinitive' (the right-thinking word in the ESOL world is "to-infinitive")... and "stop + gerund*"):

</pre_script>

Exofarts?

Good old Cambridge. Researchers there have really got their finger on the pulse. Theyve been looking at hycean planets ...

<glossary>

HYCEAN is a porte-manteau of hydrogen and ocean. Wikipedia calls them
"a hypothetical type of planet that features a liquid water ocean underneath a hydrogen-rich atmosphere", but the 'hypothetical' is based on a 2023 paper. I admit I haven't been keeping up with the literature....: 

</glossary>

...::

Astronomers have detected the most promising signs yet of a possible biosignature outside the solar system, although they remain cautious.

Using data from the James Webb Space Telescope (JWST), the astronomers, led by the University of Cambridge, have detected the chemical fingerprints of dimethyl sulfide (DMS) and/or dimethyl disulfide (DMDS), in the atmosphere of the exoplanet K2-18b, which orbits its star in the habitable zone.

On Earth, DMS and DMDS are only produced by life, primarily microbial life such as marine phytoplankton. While an unknown chemical process may be the source of these molecules in K2-18b’s atmosphere, the results are the strongest evidence yet that life may exist on a planet outside our solar system.

Original paper 

The article goes on:

The observations have reached the ‘three-sigma’ level of statistical significance – meaning there is a 0.3% probability that they occurred by chance. To reach the accepted classification for scientific discovery, the observations would have to cross the five-sigma threshold, meaning there would be below a 0.00006% probability they occurred by chance.

Well I'll give them the benefit of the doubt and turn a blind eye to that < .001% chance that they've got it wrong. But K2-18b is 124 light years away. In other words, a few years after my great grandfather Donald celebrated the birth of his son Archibald, the light that we see today from K2 (which I thought was a mountain anyway) had just set off.

Given that K2-18b is an exoplanet (a planet orbiting a star other than the sun) and the chemicals mentioned ("dimethyl sulfide (DMS) and/or dimethyl disulfide (DMDS)") are on Earth produced only by life, the exact provenance of these exofarts doesn't really bother me anyway.

And finally

I've noticed of late that something's been happening to "it's your fault", especially during the 21st century. I suppose it's down to the growing unfashionability of the idea of "fault": all this judgmental stuff isn't really politically correct. 

This Ngram from Google Books shows that in the year 2000 "it's your fault" was more than 12 times more common than "it's on you" and "it's down to you" combined, but by 2021 it was less than 7 times more common than combined sightings of the millennial arrivistes – still a good deal more common, but not nearly as much. 

That's all she wrote.

b

Summit or nuffink

In 2008 the New York Times published this report:

RIO de JANEIRO — Marina Silva, the environmental minister who resigned this week, blamed “stagnation” in the government for her decision at a news conference on Thursday and acknowledged that governors in frontline Amazon states were pressing the president to rescind measures intended to check deforestation.

“There were questions from some governors about those measures, and they couldn’t be relaxed,” Ms. Silva said.

“It is crucial that we preserve the advances we have made, it is crucial that we don’t take a step backwards,” she said.

Her resignation on Tuesday shocked the international environmental community, which saw Ms. Silva, a former rubber tapper, as a bulwark against deforestation of the Amazon.It also surprised the government of President Luiz Inácio Lula da Silva, who came into office in 2003 being hailed as his country’s first “green” president (HD: my emphasis).

'Green', eh?  Well, judge for yourself. After various comings and goings, Lula is back in power, and hosting the 50,000 environmentally-aware people who will be attending COP30 in Belém later this year.

In preparation for this event he is at last building a four-lane highway, the Avenida Liberdade, which has been on various drawing boards since 2012. It took COP, "a COP in the Amazon, not a COP about the Amazon" to use Lula's expression, to get 'spades in the ground' as politicians tend to say; 'spades in the ground' sounds a lot better than 'chainsaws in the forest'.

But a few weeks ago the BBC reported on an atrocity done on the Amazon in the name of climate activism.

A new four-lane highway cutting through tens of thousands of acres of protected Amazon rainforest is being built for the COP30 climate summit in the Brazilian city of Belém.

It aims to ease traffic to the city, which will host more than 50,000 people - including world leaders - at the conference in November.

The state government touts the highway's "sustainable" credentials, but some locals and conservationists are outraged at the environmental impact.

The Amazon plays a vital role in absorbing carbon for the world and providing biodiversity, and many say this deforestation contradicts the very purpose of a climate summit.

 

Along the partially built road, lush rainforest towers on either side - a reminder of what was once there. Logs are piled high in the cleared land which stretches more than 13km (8 miles) through the rainforest into Belém.

Diggers and machines carve through the forest floor, paving over wetland to surface the road which will cut through a protected area.

Source

The article is long and well worth reading. It discusses the people and animals (and ecosystems) that will be affected, and the livelihoods that will be disrupted (or, perhaps, ended).

<autobiographical_note>
Two weeks ago I went with MrsK to a planning meeting about a development that has a kind of relevance (although of a rather bathetic kind) to this story. When I moved to this part of the world forty years ago it struck me that locals insisted, rather quaintly, it seemed to me, on referring to Spencers Wood as 'the Village'. 'Strip development' (along the roads, leaving the green bits relatively unencroached-on) had begun in the previous thirty-odd years, with 'infill' being done only between one house and another.

Recently, though, 'infill' has had a makeover. It is now a matter of filling in the space between one strip of houses along one road, and another strip of houses on  an adjacent road. 'The Village' is becoming a suburb of Greater Reading.

<detail>
The meeting was to invite comments on an application for Outline Planning Permission for  a development of a few hundred new houses (not tens of thousands, like last time). I did comment, for what it's worth. Here is my original comment, not the ill-formatted and incomplete version that the council's SQL server (don't ask – it's a computer thing). The Lord alone knows what's going to happen when Wokingham discovers AI.

I object to this application - both in detail and in principle. It proposes a development with no credible basis (but with lip-service paid to infrastructure of course) and it invites members of the public to waste their time deciphering and criticizing a plan that is obviously incomplete. All the comments show how useless this is as a plan: the comments from Thames Water are particularly apposite: they have tried and failed to get the developers to address the plan's many obvious flaws. It is not a 'plan' at all in any useful sense. The developer obviously sees Outline Planning Permission as an irritating hoop to be jumped through and has chosen to insult the intelligence of its readers by trying to pass off a half-baked and ill thought through fever dream (sic. ...

<sic_note>
I initially put 'pipe dream' but dismissed it as a cliché, only to replace it with another – more modern but less apposite. Sorry. I'll try harder next time. But the 'plan' didn't inspire greater attention to detail. </sic_note>

...). 

</detail>

</autobiographical_note> 

Last Word

For a long time I've been wondering about the difference (if any....

<note degree_of_prescription= "0"> 
I'm really not saying 'there's no difference'. I'm just saying I can't discern  one. My grandfather certainly thought crispy was an abomination, but if I were more of a foodie I might be able to discern a difference. If there is one, I'd like to find out about it.
</note>

...) between crisp and crispy.

For some reason I've taken 1951 as a baseline for this Ngram ...

 <headslap>

(I've just thought –  my date of birth may have something to do with it). 

 </headslap> 

In 1951 crisp was more than 40 times as common as crispy; in 2022 crispy was 10 times more common than that  –   still less common, but only ¼ less common than its shorter relative. 

It now seems to have plateaued.

Basta😉 (More than...)

b

Drill baby drill

At the end of last month, an article in New Scientist  announced a research project that, if it goes ahead ( I wonder if USAID is involved in the funding somewhere down the line), will investigate a little-understood...

<understatement>
(ie NOT understood)
</understatement>

... source of oxygen: 

 

Dark oxygen: New deep sea expedition to explore mysterious discovery

The shock discovery that metallic nodules could be producing oxygen in the deep sea made headlines last year – now the team behind it is launching a new project to confirm and explain the findings

Manganese nodues

Marine scientists who made headlines last year with their discovery that deep sea nodules could be producing “dark oxygen” are embarking on a three-year research project to explain their findings.

Amid swirling controversy over their research, project lead Andrew Sweetman at the Scottish Association for Marine Science says he hopes the new scheme will “show once and for all” that metallic lumps of rock are sources of deep sea oxygen and start to explain how the process is working. “We know that it’s going on, and what we need to now do is show it again, and then really start getting at the mechanism,” he says.

Predictably, potential exploiters are queueing up to destroy the millennary bounty before there's a chance to understand it; it stands to reason, given the splendid job we've done turning Africa into a desert.

In a recent BBC radio programme – which I've tried and failed to find again, using BBC Sounds' pitiful indexing tool...

<tangent>
Honestly, it's worse than useless. They pollute the airwaves with repeated claims about their unrivalled back-catalogue, but the chances of finding anything in it are nugatory (excuse my French, it's a bit like nougat). Perhaps it's just a matter of time, and the programme was just too recent to have been properly indexed yet.
</tangent>

 ... –  Daron  Acemoglu...

<parenthesis>
(Nobel Laureate economist, recognized for his contribution to research into the link between automation and inequality)
</parenthesis>

...was talking about AI, and  different possible approaches: the Musk/Trump unregulated expensive and wasteful pro-inequality model ...

<half-remembered-quote>
'If China didn't exist the Silicon Valley elite would have to invent it.' This was said in an interesting comparison between AI-development amd the Manhattan Project..
</half-remembered-quote>

...  amd the more realistic and attainable regulated version.. If I find the programme, expect an update (but don't hold your breath).

L'envoi

This term's concert (HD 2025: mentioned in a recent post)... reminded me of this old post of mime...

<autobiographical_note type="mea maxima culpa">
(back in the halcyon days before Blogger: issued a 'much-improved' update of their indexing tool that made it virtually impossible to use, so that the word-cloud of index tags covers the first nine-odd years of this blog, but not the latest five. I should really do something about this (like hide the whole thing on the grounds of diminishing currency), but feel a strange attachment to the earlier stuff.
</autobiographical_note>

 ...which referred to a television programme about Alan Bennet's Diaries:

<prescript>
The television programme was loosely based on an edition of Private Passions, notable (to my hyper-sensitive – not to say anal – ear) for Michael Berkeley's mis-quoting of the words he had just heard (from The Dream of Gerontius): "Softly and gently, dearly ransom’d soul". He said "dear departed soul".  Come to think of it, it may not be a misquote but a quotation from elsewhere in the text, made to sound like a misquote because of the editing. He surely can’t be that cloth-eared? (Though, come to rethink of it, the angel, in the Celestial Arrivals Lounge, surely wouldn't have addressed Gerontius as departed ; he'd only just got there, for Heaven's sake.)

The collocation “departed soul” is a pretty strong one; and the syllable-count and stress pattern are right ... But it makes dear define soul, whereas in the original – by John Henry Newman –  dearly modifies ransomed.

<autobiographical_note type="hair-splitting">

A lot of ransoming goes on in Christianity. In the second line of the version of “O come O come Emmanuel” that I learned at my mother’s knee (which was never far from Aunty Katy’s, genuflecting away like billy-o,...

<digression>
(a coincidentally – I didn't know until I checked the spelling – but strangely appropriate word,  given one of the possible derivations of the word; as The Phrase Finder says,

...Alternatively, the derivation is said to be from Joseph Billio, the zealous 17th/18th century Puritan preacher. Billio preached at the United Reformed Church in Market Hill, Maldon, Essex, in and around 1696. He was an enthusiastic 'hellfire and damnation' preacher and, given his name and reputation, ought to be a serious contender as the source of the phrase. They are certainly convinced in Maldon, and it must be true - they have a plaque to prove it. 

                    But, as I was saying, genuflecting....)
                    </digression>

...as only knees can [that’s one for the etymologists]) was And ransom captive  Israel. In the  C of E-preferred version I have sung since then, the words of that line are Redeem thy cap/tive Israel. Wha...? Israel's not Emmanuel's captive  – not guilty, yer 'Onner  –  it's Pharaoh's captive. Israel was (in the 15th-century, when the carol surfaced in France) a metaphor for Christendom, and in the words of Elgar's angel, the ransom (the price paid for redemption) was dear (in the expensive sense): the soul may be dear to some people, but the point is that it was dearly ransomed.
</autobiographical_note>

</prescript> 

Anyway, here's the lowdown:

It feels like it's going to be a brilliant occasion, in a beautiful setting. Don't miss it (or the Earlybird offer – you've got just under a week to save a fiver!)

b

 

 

Moreish

 Last July I wrote (here) about Ultra Processed Food (UPF), a concept introduced by Dr Carlos Monteiro at a conference in São Paolo in 2010. My post looked at the difference....

<conspiracy-theorist type="moi?">
(in sub-editorial tone, not in substance)
</conspiracy-theorist>

... between a Medscape article about UPF and the Portuguese article it was a translation of...

<sic>
(and if you wanted me to write '... of which...', see my  earlier posts, passim. In brief there are silly pseudo-grammatical 'rules' that you have to break whichever way you deal with a subordinate clause that contains a prepositional verb: 

• 'a shibboleth about which I have written' ("should" be 'that I have written about' – that in a defining subordinate clause" insists Miss Thistlebottom

<tangent> 
(and Microsoft Word's accursed grammar checker: if you want something really fouled up, use a computer...

<meta_tangent>
I'm reminded of Amol Rajan's description [on this week's Today Podcast] of social media as 'the industrialization of confirmation bias'.
</meta_tangent>

...).
</tangent>

• 'a shibboleth that I have spoken about' (but that leaves a preposition at the end of the sentence, breaking another silly 'rule')   

As I have written before, you're damned if you do and damned if you don't. and I've chosen my path to perdition.)

</sic>

In the third of the Royal Institution Christmas Lectures one of the guests...

<parenthesis>
(Dr Chris van Tulleken, cited in that July post, presented the three lectures, but he had  a dozen or more guests, and I didn't note all their names.) 

</parenthesis>

... concocted a soft drink, saleable in today's market, which he said was dominated by products that trade on the difference between WANTING and LIKING, starting from the predictable base of water.  The first additive was six teaspoonsful of sugar...

<not_THAT_easy> 
I know I know I know. They use less fattening sweeteners. This was just a demonstration of a principle. Food scientists use cheaper/less wholesome ingredients and mask any undesirable side effects by engineering with food-like additives.
</not_THAT_easy>

... which the guinea-pig (Dr VanT) found unpleasantly sweet. The next additive was citric acid, to counteract the sweetness. Again the doctor found a reason for disliking the improved drink, and again the guest added something else. This went on for several more steps, with the additives getting more and more arcane,  and their net result was a drink that started out with an excessive amount of sugar (an amount that the body was wise enough to reject), but that had that natural feeling of satiety cloaked; the body wanted more and took it before it had a chance to realize that the sweetness was not a healthy amount.

According to the OED, English has had a word for this since the late seventeenth century, but had scarcely any use for it until the late twentieth century, when discoveries – and exploitation of those discoveries – popularized it: moreish. In fact, it was so rarely used that unlike many other words (most? – Discuss) it has retained only one meaning for over three centuries:

(Don't bother clicking on any apparent links; this is just a dumb screenshot.)

And this Ngram from Google Books shows that for the first 290 years of its 3⅓ centuries' life the word was flying below the radar, virtually invisible; then it rocketed up (to use a slightly less inappropriate [but pleasingly aeronautical] metaphor:

While researching this post I was struck (or do I mean stricken?) by a phrase in the Medscape article quoted in my July post  that referred to UPF's 'hyperpalatability and high caloric density'. I suppose 'hyperpalatability' just means 'moreishness'; and 'high caloric density' means 'hidden ability to make you eat more than is good for you'.

<Newspeak_reflection>
The apparatchik that strung together those words (I'm not sure such a person would deserve the term 'writer') was, in Newspeak terms, a 'doubleplusgood duckspeaker'.
</Newspeak_reflection>

At the risk of confirming my nearest and dearests' feeling that I always see the dark side of the Sun, I'm inclined to conclude that there's a red flag here: if you describe something as 'moreish' it's almost certainly UPF.

That's quite enough.

b

Not letting the facts interfere with a good headline

'Well don't say I didn't warn you', as Stephen Hawking might have said. What he actually said was this:

“Success in creating effective AI, could be the biggest event in the history of our civilization. Or the worst. We just don’t know. So we cannot know if we will be infinitely helped by AI, or ignored by it and side-lined, or conceivably destroyed by it,” ...

“Unless we learn how to prepare for, and avoid, the potential risks, AI could be the worst event in the history of our civilization. It brings dangers, like powerful autonomous weapons, or new ways for the few to oppress the many. It could bring great disruption to our economy.”   

         Source 

Last week Apple put their foot in it, in a mistake that, while essentially trivial, had  huge implications for journalism. This extract from a  BBC report of the (ridiculous) event ends 'Apple has made no comment.' But that doesn't mean they did nothing. In my search for traces of the story I was often stymied by 404 errors, and I suspect there has been an attempt at a web-wide clean-up attenpt, thankfully resisted by the BBC

A major journalism body has urged Apple to scrap its new generative AI feature after it created a misleading headline about a high-profile killing in the United States.

The BBC made a complaint to the US tech giant after Apple Intelligence, which uses artificial intelligence (AI) to summarise and group together notifications, falsely created a headline about murder suspect Luigi Mangione.

The AI-powered summary falsely made it appear that BBC News had published an article claiming Mangione, the man accused of the murder of healthcare insurance CEO Brian Thompson in New York, had shot himself. He has not.


Now, the group Reporters Without Borders has called on Apple to remove the technology. Apple has made no comment.

T

 

Screen grab of AI's Grossest Hour
(not its finest)

The BBC was not a;lone:

 

Reporters Without Borders (RSF) said this week that Apple's AI kerfuffle, which generated a false summary as "Luigi Mangione shoots himself," is further evidence that artificial intelligence cannot reliably produce information for the public. Apple Intelligence, which launched in the UK on December 11, needed less than 48 hours to make the very public mistake. 

"This accident highlights the inability of AI systems to systematically publish quality information, even when it is based on journalistic sources," RSF said. "The probabilistic way in which AI systems operate automatically disqualifies them as a reliable technology for news media that can be used in solutions aimed at the general public."

Because it isn't reliably accurate, RSF said AI shouldn't be allowed to be used for such purposes, and asked Apple to pull the feature from its operating systems.

Source

But an expert speaking last week on BBC News did not fully agree that 'AI shouldn't be allowed to be used for such purposes'. She said that there should always be a human being 'in the loop...

<tangent>
She really did say that, I'm afraid. 'Which loop?' I remember thinking at the time. When there are several people sharing information, and a boss says 'Keep me in the loop', that makes sense. It's still gross managerese, it's still a lamentable cliché, but at least it makes sense. In this case, though, an AI engine is exposed to thousands of reports and millions of possible summaries, and regurgitates a patent lie. Where's the loop in that?
</tangent>

... '

The Byte reported the mistake thus:

"I can see the pressure [HD: on? for? to get?] getting to the market first, but I am surprised that Apple put their name on such [a....

<tangent>
With acute regret I have to concede that the speaker may have meant to omit the article – he was, after all, a 'professor in media policy' (whatever that is).
</tangent>

...?] demonstrably half-baked product [HD's emphasis; I just applaud the phrase]," Petros Iosifidis, a professor in media policy at City University in London, told the BBC. "Yes, potential advantages are there — but the technology is not there yet and there is a real danger of spreading disinformation."

However, this danger is one that's fundamental of [HD:sic...

<tangent> 
'Oh dear', as my grandson might say. This bit of writing really leaves a lot to be despaired. Meanwhile back at that article...
</tangent> 

...generative AI, and not just Apple's flavor of it. AI models routinely hallucinate and make up facts. They have no understanding of language, but instead use statistical predictions to generate cogent-sounding text based on the human writing they've ingested  

 And finally..

 On 22 March 2025 my choir  is joining forces with Bracknell Choral Society. for a performance of The Dream of Gerontius:

<autobiographical-note>
The first time I sang this extraordinary piece my mother (whom saints preserve [and they better had]) was in the audience.
</autobiographical-note>

You've got just over a month (as from publication date) to snap up the Earlybird deal.

b