This post is about a conversation I had with Google Images and his AI assistant, a conversation which took place when I was not logged in. Which may well have affected how much effort Google put into the conversation.
It was triggered by a dug-over patch appearing on a verge adjacent to the roundabout where Meadway crosses West Hill Avenue, snapped above, where BH thought that there had been a mature oak tree, a cause for concern for the occupant of the house behind.
I had no recollection of the tree in question, despite the fact that I usually pass the spot several times a week. Twice a day when I was still working, but that was twenty years ago.
The conversation was very like one with a human. That is to say rather disorganised and quite a few wrong turnings and mistakes.
Also like a television detective drama – say Morse – in which the detective concocts a plausible story around the few facts at his disposal. And sometimes the facts are not facts at all and he has to concoct a new story.
Long summary of the conversation
The conversation with Google Images and his AI assistant (G). It started on the morning of Tuesday, 28th July. In what follows, my inputs are normal text, while Googles replies are summarised as indents.
Round one
[A winter snap of the tree, lifted from Street View, without leaves. Such images are much lower resolution that those which come straight from my telephone, but still look clear enough to the naked eye. A cropped version of the snap above].
A road side tree near Clay Hill Green in Epsom. Said to be an oak but it does not look much like an oak to me.
• G offers three reasons why the tree look more like an ash than an oak
• He says that there are a lot of oak trees in Epsom, but there are plenty of others too. And this one is an ash.
Since the camera-man went by, the tree has been felled. If it was an ash tree, might this have been one of the two ash diseases which I believe are on their way?
• G tells me all about ash dieback (now) and emerald ash borer (to come)
• Epsom & Ewell Council know all about ash dieback
The tree does not appear on the map of ash trees - which looks as if someone has tried to enumerate every ash tree in the borough - in the appendix to the action plan. But I imagine there are going to be gaps in such an enumeration.
• G gives me various good reasons why this might be. One being that Epsom Borough does not do road-side trees, which are mostly a Surrey matter
I have my own ash tree, maybe thirty five years old with a fair bit of dead wood in its lower parts. But last time I sent a picture to the Forestry Commission, they said no problem. I shall take another, more careful look.
• A digression. G gives various tips on diagnosing ash dieback.
How do I save this conversation for future reference? Copy and paste?
• Cut and paste is the way to do it. While Gemini, at least when you are paying, keeps the history for you. At least most of the time
Round two, a bit later in the afternoon
As it happens, from a different viewpoint, Street View offers a different, much more recent image, with leaves, and with no dead wood visible. Can I upload that without you losing the thread?
• Yes
The new image, dated October 2022. The first one was dated March 2018.
• G changes tack with the new leafy image, a cropped version of that above. The tree is now an oak after all, specifically Quercus robur. This despite the leaves not being resolvable in the image from Street View. Low resolution compared, for example, with my telephone
• He goes on to put the now-missing tree on the edge of Clay Hill Green and the proposed Allonby development there
This is the sort of thing that the Epsom Civic Society would have taken an interest in. I have had a quick look, without success. Are you able to 'read' their website?
• G says that he can, and goes on to tell me about their activities. With this particular tree being covered as part of their response to Allonby
This tree is on the roundabout at the junction of Meadway and West Hill Avenue. I think the development you are talking about was a bit further to the south, adjacent to Clay Hill Green itself, complete with listed wall.
• G says that this makes it a Surrey matter and is probably all mixed up with damage to utilities. In which case, the tree can be cut down with much less bother
• He points the Civic Society being concerned about the loss of roadside trees in this area
Round three, a bit later still in the afternoon
Going back to Allonby, following the link you gave, I read that the council rejected the application for development in January 2022. And I have yet to see any sign of action at the site. Just a bit of wall which fell down and has not yet been reinstated.
• G acknowledges his error, without attempting to explain how he came to make it. He then picks up on heritage walls
I am not that keen on preserving old walls. The council should have more important things to do with its money - and with ours!
• G is a happy to go along with anti-wall too!
THE END
Short summary of the conversation
During round one, G locks onto his identification of the tree as an ash tree. He is quite good on on ash dieback in general and on the Council’s response in particular.
At the start of round two, he switches to a specific species of oak.
I did not ask, but he does not comment on how you can sometimes get quite different views of the same thing in Street View. Although I can see that checking with Street View without a more precise location might have been difficult.
But he does lock onto the Allonby development a little way down the road. When I point out his error, his changes tack without offering any explanation of how he came to make the error.
In the snap above, the tree can be seen to the right of the upper spot. The development was to be under the lower spot.
In this context, gmaps was much more useful than OS. Maybe they do satellite view too, but I have not found it.
In round three, I point out his error about the status of this development. He can put his hand on the relevant document and goes on to pick up on my throwaway about heritage walls. Again, without offering any explanation.
The parable of the countryman
I vaguely remember my father, who was brought up in the country between Huntingdon and St. Ives, telling me that a countryman could identify a tree, say an isolated hedgerow tree on the other side of the field, by its general shape and appearance. No need to inspect the leaves, let alone the buds, fruits or flowers, or for any of that fancy stuff that botanists go in for.
It occurs to me now that his countryman might not be so much cleverer than Google Images after all.
First, the countryman probably worked on a farm and probably knew all the trees on and around his farm. The general shape and appearance of the tree, winter or summer, was a help, an aide-memoire, but he could pull up the identification from a catalogue in his memory.
And yes, his brain had digested all this information and would have a stab at a strange tree, a very confident sounding stab. But how often would you actually check him? How often would this be convenient or possible? Maybe he actually got it wrong quite often.
So maybe, the botanists are right. You cannot be sure of a tree’s genus – never mind – species from its general appearance from a distance, any more than you can be sure of its age from its girth, let alone its height. These inputs just do not tell one enough about the tree’s unique life history.
On a quick glance, for example, one is quite likely to call a pollard oak by a river a willow. Pollard willows being a conspicuous feature of the riverine landscape, certainly in Huntingdonshire.
Photographical aside: not much dead looking wood in the winter snap, which one might expect in a diseased ash. While in the summer snap, the tree looks rather bigger – and perfectly healthy. To be fair, it is also four years older.
Botanical aside, turned up by Bing: ‘Oaks are in the genus Quercus which is in the Beech family. There are more than 600 species worldwide. There are two native oaks in Britain, the English Oak and the Sessile Oak. Three more oaks are shown here but there are many more species in parks, urban settings or botanical gardens in Britain. These five species show many of the key characteristics of the oak genus. Oak tree identification – English, Sessile, Turkey have lobed leaves, Holm and Cork have evergreen leaves, all 5 have male catkins and acorns’.
Conclusions
G makes a positive error in his identification of the leafless tree. He is equally positive about the leafy tree. I have not done the necessary work to confirm his second attempt – although I dare it was an oak, if not necessarily the specific species suggested.
He jumps to conclusions about the location of the tree.
He gets the planning status of the Allonby development wrong, despite the right answer being readily available to him.
He was good on the working of the ash diseases, the Civic Society and the Councils. Matters which could be well dealt with in words, without needing to get into the tricky interactions between words and pictures.
Maybe he would have done better had I been logged in. But one of his lesser efforts, nonetheless.
PS: by way of contrast, he was able to put me right without fuss on working on two documents at once in Word today.
References
Reference 1: https://www.surreycc.gov.uk/land-planning-and-development/interactive-map. A clever map, but I did not find it helpful in this context.
Reference 2: https://www.treeguideuk.co.uk/oaks/. Helpful on distinguishing the various oaks commonly found in England.
Reference 3: https://www.epsom-ewell.gov.uk/sites/default/files/documents/residents/access-nature/trees-epsom-ewell/Ash%20Dieback%20Action%20Plan%202023%20%2B%20Map.pdf. Council on ash.
Group search key: aisk.


















.jpg)











.jpg)
.jpg)


.jpg)



















