On 2025-02-15, Roger Hayter <roger@hayter.org> wrote:
Someone sent me a WhatsApp message saying inter alia "Sorry I don't live
nearer! Otherwise I could have <performed useful task>"
I am using Mac OS 15.3 which claims to have some AI feature, though I have >> avoided all offers to explicitly enable it. The OS puts notifications of
WhatsApp messages at the top of the screen, which in the past have been the >> first two or three lines of the actual message. On this occasion I got a
notification the start of which was in italics, which I have later inferred >> must be an AI summary, though I originally assumed that the message
must have been written in italics and that this was the unaltered text.
The beginning of the message said "I apologise for not being closer ..."
This is so different in meaning, at first glance implying a social or
emotional closeness rather than a geographical one, that it could lead
someone to totally misinterpret the message with embarrassing
consequences. If these consequences cause significant harm could one
sue Apple? They do attempt to summarise human messages, so I do not
see how they can avoid responsibility for foreseeable consequencies.
With the T&Cs, presumably.
I know one iteration of Apple AI has already been criticised for
reversing the meaning of news headlines when summarising them. Perhaps
Apple are just not very good at AI - or could there be a malign
purpose behind this?
You're assuming that "better AI" exists somewhere outside Apple.
Someone sent me a WhatsApp message saying inter alia "Sorry I don't live nearer! Otherwise I could have <performed useful task>"
I am using Mac OS 15.3 which claims to have some AI feature, though I have avoided all offers to explicitly enable it. The OS puts notifications of WhatsApp messages at the top of the screen, which in the past have been the first two or three lines of the actual message. On this occasion I got a notification the start of which was in italics, which I have later inferred must be an AI summary, though I originally assumed that the message
must have been written in italics and that this was the unaltered text.
The beginning of the message said "I apologise for not being closer ..."
This is so different in meaning, at first glance implying a social or emotional closeness rather than a geographical one, that it could lead someone to totally misinterpret the message with embarrassing
consequences. If these consequences cause significant harm could one
sue Apple? They do attempt to summarise human messages, so I do not
see how they can avoid responsibility for foreseeable consequencies.
I know one iteration of Apple AI has already been criticised for
reversing the meaning of news headlines when summarising them. Perhaps
Apple are just not very good at AI - or could there be a malign
purpose behind this?
On 2025-02-15, Roger Hayter <roger@hayter.org> wrote:
Someone sent me a WhatsApp message saying inter alia "Sorry I don't
live nearer! Otherwise I could have <performed useful task>"
I am using Mac OS 15.3 which claims to have some AI feature, though I
have avoided all offers to explicitly enable it. The OS puts
notifications of WhatsApp messages at the top of the screen, which in
the past have been the first two or three lines of the actual message.
On this occasion I got a notification the start of which was in
italics, which I have later inferred must be an AI summary, though I
originally assumed that the message must have been written in italics
and that this was the unaltered text.
The beginning of the message said "I apologise for not being closer
..."
This is so different in meaning, at first glance implying a social or
emotional closeness rather than a geographical one, that it could lead
someone to totally misinterpret the message with embarrassing
consequences. If these consequences cause significant harm could one
sue Apple? They do attempt to summarise human messages, so I do not see
how they can avoid responsibility for foreseeable consequencies.
With the T&Cs, presumably.
I know one iteration of Apple AI has already been criticised for
reversing the meaning of news headlines when summarising them. Perhaps
Apple are just not very good at AI - or could there be a malign purpose
behind this?
You're assuming that "better AI" exists somewhere outside Apple.
Someone sent me a WhatsApp message saying inter alia "Sorry I don't live nearer! Otherwise I could have <performed useful task>"
I am using Mac OS 15.3 which claims to have some AI feature, though I
have
avoided all offers to explicitly enable it. The OS puts notifications of WhatsApp messages at the top of the screen, which in the past have been
the
first two or three lines of the actual message. On this occasion I got a notification the start of which was in italics, which I have later
inferred
must be an AI summary, though I originally assumed that the message must
have
been written in italics and that this was the unaltered text.
The beginning of the message said "I apologise for not being closer ..."
This is so different in meaning, at first glance implying a social or emotional closeness rather than a geographical one, that it could lead someone
to totally misinterpret the message with embarrassing consequences. If
these
consequences cause significant harm could one sue Apple? They do attempt
to
summarise human messages, so I do not see how they can avoid
responsibility
for foreseeable consequencies.
I know one iteration of Apple AI has already been criticised for
reversing the
meaning of news headlines when summarising them. Perhaps Apple are just
not
very good at AI - or could there be a malign purpose behind this?
Roger Hayter wrote:
Perhaps Apple are just not very good at AI
today's the day that Google start going live with their digital fingerprinting of users' machines to target advertising more directly.
https://www.bbc.co.uk/news/articles/cm21g0052dno
Nick wrote:
Roger Hayter wrote:
Perhaps Apple are just not very good at AI
today's the day that Google start going live with their digital
fingerprinting of users' machines to target advertising more directly.
https://www.bbc.co.uk/news/articles/cm21g0052dno
"By allowing fingerprinting, Google has given itself - and the
advertising industry it dominates - permission to use a form of
tracking that people can't do much to stop," said Martin
Thomson, distinguished engineer at Mozilla"
That might be reassuring, except Mozilla has declared their own future
lies in advertising and AI ... if Google keep throwing them enough crumbs.
On Sun, 16 Feb 2025 14:13:39 +0000, Andy Burns <usenet@andyburns.uk>
wrote:
Nick wrote:But at least by eschewing Alphabet, Apple and Microsoft operating
Roger Hayter wrote:
Perhaps Apple are just not very good at AI
today's the day that Google start going live with their digital
fingerprinting of users' machines to target advertising more directly.
https://www.bbc.co.uk/news/articles/cm21g0052dno
"By allowing fingerprinting, Google has given itself - and the
advertising industry it dominates - permission to use a form of
tracking that people can't do much to stop," said Martin
Thomson, distinguished engineer at Mozilla"
That might be reassuring, except Mozilla has declared their own future
lies in advertising and AI ... if Google keep throwing them enough crumbs. >>
systems, browsers and search you are giving yourself a fighting
chance.
Nick
Now that everybody - Apple Google, Meta, Microsoft et al - are
aggressively pushing stuff many users don't want, if one hasn't already, maybe now is the time to consider a Linux OS. It used to be awfully
geeky but it isn't that way any more. To steal a slogan from Apple and
apply it to Linux: "It just works."
..and today's the day that Google start going live with their digital >fingerprinting of users' machines to target advertising more directly. >https://www.bbc.co.uk/news/articles/cm21g0052dno
On Sun, 16 Feb 2025 14:00:22 +0000, nick <nickodell49@yahoo.ca> wrote:
..and today's the day that Google start going live with their digital
fingerprinting of users' machines to target advertising more directly.
https://www.bbc.co.uk/news/articles/cm21g0052dno
Good.
Yes, really. Good. It might actually help reverse the trend towards websites being either paywalled or completely smothered in intrusive adverts.
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/
Mark
Yes, really. Good. It might actually help reverse the trend towards websites being either paywalled or completely smothered in intrusive adverts.
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/
On 16/02/2025 22:12, Mark Goodge wrote:
Yes, really. Good. It might actually help reverse the trend towards websites >> being either paywalled or completely smothered in intrusive adverts.
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/
You might want to include cloudflareinsights on your cookies page
(although the cross origin request is being blocked by firefox anyway).
On Mon, 17 Feb 2025 09:22:12 +0000, Nick Finnigan <nix@genie.co.uk> wrote:
On 16/02/2025 22:12, Mark Goodge wrote:
Yes, really. Good. It might actually help reverse the trend towards websites
being either paywalled or completely smothered in intrusive adverts.
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/
You might want to include cloudflareinsights on your cookies page
(although the cross origin request is being blocked by firefox anyway).
Good point; I wrote that page before I switched the site to Cloudflare.
On 17/02/2025 10:40, Mark Goodge wrote:
On Mon, 17 Feb 2025 09:22:12 +0000, Nick Finnigan <nix@genie.co.uk> wrote: >>
On 16/02/2025 22:12, Mark Goodge wrote:
Yes, really. Good. It might actually help reverse the trend towards websites
being either paywalled or completely smothered in intrusive adverts.
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/
You might want to include cloudflareinsights on your cookies page
(although the cross origin request is being blocked by firefox anyway).
Good point; I wrote that page before I switched the site to Cloudflare.
https://en.wikipedia.org/wiki/Cloudflare
"Internet security
[snip]
"The product, instead of presenting a visual CAPTCHA for the user to
solve, automatizes the verification process by conducting
JavaScript-based checks inside the browser to determine whether the user
is a real person or an automated entity."
Sounds dodgy. How does it really work?
On Sun, 16 Feb 2025 14:00:22 +0000, nick <nickodell49@yahoo.ca> wrote:
..and today's the day that Google start going live with their digital >>fingerprinting of users' machines to target advertising more directly. >>https://www.bbc.co.uk/news/articles/cm21g0052dno
Good.
Yes, really. Good. It might actually help reverse the trend towards
websites
being either paywalled or completely smothered in intrusive adverts.
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/
On Sun, 16 Feb 2025 22:12:44 +0000, Mark Goodge wrote:
https://www.markgoodge.com/2024/10/why-the-web-is-getting-worse/Thanks for writing that, Mark.
While I was reading your article I found I had Gillian Welch singing in
my head:
<https://www.youtube.com/watch?v=31qwSm3chn4>
On 15/02/2025 23:15, Roger Hayter wrote:
Someone sent me a WhatsApp message saying inter alia "Sorry I don't live
nearer! Otherwise I could have <performed useful task>"
I am using Mac OS 15.3 which claims to have some AI feature, though I have >> avoided all offers to explicitly enable it. The OS puts notifications of
WhatsApp messages at the top of the screen, which in the past have been the >> first two or three lines of the actual message. On this occasion I got a
notification the start of which was in italics, which I have later inferred >> must be an AI summary, though I originally assumed that the message must have
been written in italics and that this was the unaltered text.
The beginning of the message said "I apologise for not being closer ..."
This is so different in meaning, at first glance implying a social or
emotional closeness rather than a geographical one, that it could lead someone
to totally misinterpret the message with embarrassing consequences. If these
consequences cause significant harm could one sue Apple? They do attempt to >> summarise human messages, so I do not see how they can avoid responsibility >> for foreseeable consequencies.
I know one iteration of Apple AI has already been criticised for reversing the
meaning of news headlines when summarising them. Perhaps Apple are just not >> very good at AI - or could there be a malign purpose behind this?
To answer the question in your subject:
What does your contract with the suppliers say on the matter?
What consideration did you give to said suppliers in exchange for the AI specifically, (rather than it being bundled for free on a device you purchased)?
Would turning off the notifications not be a more appropriate remedy
than issuing against said supplier? [^1]
Regards
S.P.
[^1] Go to Settings > Notifications. Scroll to the list of Apps and
find WhatsApp then select it. Scroll down to the bottom and tap the
toggle next to "Summarize Notifications" to turn it off.
On 18/02/2025 10:45, Roger Hayter wrote:
On 18 Feb 2025 at 10:33:08 GMT, "Simon Parker" <simonparkerulm@gmail.com>
wrote:
Would turning off the notifications not be a more appropriate remedy
than issuing against said supplier? [^1]
I find the notifications handy. Now I know they are defective apart from the >> author of the message I can simply ignore their content
My understanding is that you can revert to the former style of summary, (where you see the first <n> characters / lines of the message), by
leaving notifications on but disabling the summarising of these notifications.
But I'm not an Appleist so I fear you'll need to play around with the settings.
snip footnote
On 18 Feb 2025 at 11:39:08 GMT, "Simon Parker" <simonparkerulm@gmail.com> wrote:
On 18/02/2025 10:45, Roger Hayter wrote:
On 18 Feb 2025 at 10:33:08 GMT, "Simon Parker" <simonparkerulm@gmail.com> >>> wrote:
snip
Would turning off the notifications not be a more appropriate remedy
than issuing against said supplier? [^1]
I find the notifications handy. Now I know they are defective apart from the
author of the message I can simply ignore their content
My understanding is that you can revert to the former style of summary,
(where you see the first <n> characters / lines of the message), by
leaving notifications on but disabling the summarising of these
notifications.
But I'm not an Appleist so I fear you'll need to play around with the
settings.
snip footnote
Just to report that when you look for the global switch for "summarizing" notifications it contains the helpful statement: "Summaries may contain errors." Clearly the fault is turning this rubbish on by default, but I thank contributor for the unsurprising advice that complaining will get me nowhere.
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 497 |
| Nodes: | 16 (3 / 13) |
| Uptime: | 29:48:47 |
| Calls: | 9,797 |
| Calls today: | 16 |
| Files: | 13,749 |
| Messages: | 6,188,690 |