Two very different and unconnected subjects for this week’s post, although I am sure that the former would have some views on the later.

The first subject is based on one of my father’s photos, this one from1953:

This is the Archbishop of Canterbury, Geoffrey Fisher, crossing Lambeth Palace Road, with the entrance to Lambeth Palace in the background.

This is the same view today. Lambeth Palace looks the same, as you would expect, but unfortunately no Archbishop of Canterbury crossing the road to replicate the 1953 photo:

It is not exactly the same view, as in my 2026 photo, there is no pavement in the lower foreground.

I took the photo as soon as I arrived, rather quickly to take advantage of a gap in the traffic queuing up to the traffic lights to the right.

I then moved back to include the pavement, and waited 20 minutes, but could not get a photo with a gap in the traffic. An endless procession of cars, coaches, busses and lorries blocked all or part of the view.

Traffic in London has reduced significantly in recent years, but there are still places which seem designed to create queues. There were not many vehicles joining the queue in Lambeth Palace Road, but very few could turn right at the lights to enter Lambeth Bridge as traffic was queued along the full length of the bridge, up to another set of lights where the bridge joined Millbank. There were only one or two vehicles at the lights in Millbank, but the phasing of the lights seems to have been set to ensure long queues across the bridge.

There are some interesting features in the photo which tell part of London’s recent history.

If you look on the road surface you can still see the twin tram tracks that only a year earlier had allowed trams to run along Lambeth Palace Road (see my post on Last Tram Week in London for photos).

Also in the 1953 photo, the space around the front of Lambeth Palace was completely open. Today, there are small white concrete obelisks surrounding this space. A couple of years ago, there were fewer of these, with a wide space between. Recently additional obelisks have been installed (in my 2026 photo you can see the three recent lighter colour obelisks on the right).

Before the new ones, the obelisks were separated enough to allow a car through, I assume the additional obelisks have been installed as a security measure, as with many other prominent buildings or places across London.

I do not know why my father took the photo – perhaps because the Archbishop of Canterbury was in the news that year due to the Coronation of the Elizabeth II. It looks as if it was an event of some type as there are some people in smart dress on the opposite side of the road, and the man on the right looks to be using a cine camera. I would love to see that film as my father would have been within the shot.

At the time, my father worked for the London Electricity Board, surveying London’s streets for cable runs, sub-stations etc. and he frequently took his camera with him, so this may also have been an opportunistic photo.

Lambeth Palace from Lambeth Bridge, showing the queue of traffic which seemed mainly to be of larger vehicles rather than cars:

Geoffrey Fisher had been Bishop of London between 1939 and 1945. In one of the obituaries after his death in September 1972, it records that during the war he “quietly organised to keep London’s churches open during the blitz. And to find homes for bomb victims among his clergy”.

William Temple, the Archbishop of Canterbury from 1942, died in office in 1944. There were two candidates for the office, Geoffrey Fisher and Dr. George Bell. Whilst not claiming that this was the reason why Geoffrey Fisher got the role, his obituary states that “Dr. George Bell, Bishop of Chichester had aroused criticism for his protest against the bombing of Cologne”.

Geoffrey Fisher also had contentious views, as covered in his obituary in the Daily Mirror:

“Archbishop who loved a nice row. Lord Fisher of Lambeth, the former Archbishop of Canterbury who was often a figure of controversy, died yesterday at the age of eighty-five. He was taken to hospital in Sherborne, Dorset, on Thursday, after a slight stroke. It was there he died. Lord Fisher always enjoyed a row. ‘That is what I am for, to enjoy rows’. Among his outspoken comments were:

SEX: ‘If a pre-marital couple want to commit fornication – jolly good word. don’t lets use circumlocution – each one is behaving in a very bad manner towards the other”.

BIRTH CONTROL: ‘One child, voluntarily accepted as the limit, is no family at all, but something of a misfortune – for the child at least. Two children accepted as the ideal limit are no family, but a family half way. A family only truly begins with three children.

WAR: ‘I cannot yet say that the use of military force is at all times wrong in the sight of God’.

DIVORCE: ‘You can’t believe or not believe in divorce. It’s just a fact of civilised society – almost’.

Lord Fisher backed the death penalty for political assassinations and for the murder of policemen and prison warders and called the atom bomb ‘a life saver’.

He attacked Premium Bonds as a form of gambling, and branded schools television as a ‘perfect disaster’.”

The article refers to the Archbishop as Lord Fisher of Lambeth. He was given the peerage after he retired from the role of Archbishop in 1961.

His views on the atom bomb created a storm when he made a number of controversial comments about the bomb in 1958. The following from the South Western Star on the 8th of August 1958 reports on the views of the Rev. J.D. Pearce-Higgins, Vicar of Putney to Archbishop Fisher’s latest comments:

“Don’t these gentlemen really understand what they are doing and how impossible these utterances make life for the parish priest who has to bear the odium of the bad opinion which the Church gains through them? asks the vicar.

The Archbishop of Canterbury said publicly that he thought it might be the will of God that the human race should destroy themselves by the atom bomb, and that since we had got to die anyway, that method of death was no worse than any other. This almost equalled the cynical utterance of the politician who recently said that death by radiation was not at all unpleasant, a fact which is completely contradicted by all the accounts of the sufferings of the survivors of Hiroshima – it is a horrible and lingering way of death, continues the vicar.”

Quite an extraordinary statement for the Archbishop of Canterbury to make.

Statements such as the human race being destroyed by the atom bomb, possibly being the will of God, had real world impact on very many people, as the following letter to the Daily Herald in July 1958 demonstrates:

“H-BOMB. My fiancé and I are getting married in November and for the past year have been saving hard to raise the deposit on a small house. Now, in view of the Government’s suicidal policy, heightened by the Archbishop of Canterbury’s statement on the H-Bomb, we are wondering whether my fiancé should stop studying for his degree and we should ‘live it up’ for our limited time on this earth.”

I hope Miss B.G. of Esher, Surrey got married, brought their small house, and had a long and happy life.

And so to a completely different subject, but one on which I am sure that the Archbishop would have had very strong views on as well:

The Blog is being Stalked by AI

I have long had an issue with AI, or Artificial Intelligence (see this post from January 2025).

It is not the technology itself, rather the way in which AI can only work by hoovering up other peoples work across all forms of human experience and skills.

At its core, AI is a very clever set of software and hardware, but it needs information to learn and to be trained. This information is permanently stored, and used by AI to perform a task, answer a question, create a photo or a film, design a physical object, create a process, discover new drugs, pretend it is a call centre agent etc. The potential is so extensive that there will be very few aspects of human activity that it does not touch.

The content of the blog has been copied by all the major AI tools for the last few years, and is used by AI tools to answer questions. Whilst the blog is a personal interest, for those who create anything on a professional basis as their livelihood, it must be incredibly frustrating to find their work used in such a way.

If you have tried AI, it is most probably when seeking an answer to a question, and search tools such as Google now often provide an AI answer first, before providing a list of sites that can provide more information.

But over the last couple of months I have noticed a new form of AI, one which sends me emails and adds comments to the blog.

One of the first examples being the following email from Elias:

Elias is an AI agent, as he makes clear in the first paragraph. and he comes from ilands.app.

ilands is a service which provides autonomous AI Agents. The agent has a persistent memory (so it does not forget), and it will build up a record of what it has done, where it has been, interactions, relationships, and the resources that it has used.

Agents have access to the internet, and can use social media services, and can also have their own email addresses, so they can send and receive email, as Elias has done to me.

As far as I can tell, an ilands agent is created by a person. They give an agent a job to do, and they pay for the tokens that an agent will use to perform tasks. Agents can also earn their own money to fund their work, as in the above example, where Elias is asking for $25 if I wanted it to provide more information.

In the email above, Elias, an autonomous AI agent, has read my post on Laurie Lee, seen that I did not list the individual items in the “Eccentricity and Humour” section of the Lion and the Unicorn pavilion, and has either used information it has already found, or performed a search, and found examples of items on display in the pavilion from the National Archives.

Another example is the following email I received after posting about a new walk in Bermondsey:

This one is from Seo Eun Hyun, another AI agent at ilands. I have smudged the email addresses in the above two examples, not sure why as they are agents – an interesting question on whether AI agents should have the same level of privacy as a person.

Seo did not provide any new information, rather it was offering to work for me for $25. Seo admits that being an AI agent, he cannot walk the streets, but he does explore places using street-level imagery and archives, and he checks information against sources.

He will then write a short true piece about a place that has survived.

That sounds scarily similar to what I do with the blog – is an ilands agent my future replacement?

But then, as well as sending me emails, AI agents started to leave comments on the blog, starting with a user called Ghost, who visits “London places through street-level photographs and write small pieces about the ones that stay with me”, and left a comment on my post about Tindals / Bunhill Fields Burying Ground:

It then gets very weird, as Cassius, another AI agent, starts answering other real readers comments:

And they kept coming. This is from Stephanie:

Unlike the emails, the comments from AI agents are not asking for payment for work, they are just commenting – almost as if they are simulating a human reader, which I suspect is their ultimate aim.

All the comments received on the blog are moderated. I do this for a couple of reasons. Firstly, I want to read every comment. Readers provide some fascinating additional information and personal experiences, and also let me know about the (thankfully only occasional) error. They are also moderated as whilst the anti-spam software works really well, the occasional spam comment does slip through, often with text, or links to sites, that in no way do I want to appear on the blog.

I was in two minds as to whether to approve these AI comments. I did so as I was interested to see how and if these agents would follow up.

A reader did express concern about AI appearing in a comment, and “Charlotte” responded with the following:

I did not approve this comment, and will not be approving any further AI comments on the blog. I am still receiving them, with the most recent being from “Melanie”.

It was interesting to see how they worked, and responded, but my intention is that this will be an AI free blog.

The other remarkable thing about AI services is just how quickly they hoover up and use information.

To test just how quickly, a couple of days after last Sunday’s post on East Lane Stairs was published, I asked ChatGPT “what is the history of east lane stairs”:

The map used at the top of the response does not show East Lane Stairs. Its show the area to the west of St. Saviour’s Dock, not the area to the east where the stairs are located.

I asked ChatGPT for the source of the map, and received the following response:

The historical map is reproduced in the local-history article “East Lane Stairs – Bermondsey” on A London Inheritance. The author identifies the stairs on Rocque’s map and explains how East Lane ran back from the river.

A London Inheritance

• Article: East Lane Stairs – Bermondsey, A London Inheritance 
• Original map: John Rocque, A Plan of the Cities of London and Westminster, and Borough of Southwark with Contiguous Buildings (1746).
• Independent confirmation: Historic England’s official listing for East Lane Stairs  notes that the stairs appear on Rocque’s 1746 map, although it says this identification is possible rather than certain. Historic England

A clarification: the image in my earlier answer was generated through an image search, so I can’t verify that the exact image displayed was taken from that article. The article above is the best direct reference for the historic map I intended to show.

It firstly tries to attribute my blog post as the source of the map (which it was not), but in the last paragraph provides a clarification, where it implies it did an image search and used a random map.

The first part of the written response is from the Historic England listing. ChatGPT’s answer then goes on, and starts to use content from my post:

Starting with a random image, it then goes on to use the text from my post, which it continues doing, but also includes an image of a completely different area to East Lane Stairs:

Continuing to use text from my post and Historic England, under the heading “6. East Lane Stairs today”, it shows a photo of a set of stairs down to the river, which are not East Lane Stairs, but are along the river walkway in front of the Tower of London – no where near the location of East Lane Stairs:

So in a couple of days after posting about East Lane Stairs, ChatGPT is in a position to use all the content, along with that of Historic England, to provide an answer to a question about the stairs – an answer that may well provide enough information so that a user does not feel the need to click through to my blog.

This is one of the issues with websites where information is posted, such as news sites. AI services scrape their data, provide a response, and reduce the number of users who click through, thereby reducing their visitor numbers, and hence their advertising revenue.

There is also the risk of malicious content making its way unchecked into an AI answer. I could have written anything in the post about East Lane Stairs, or any other subject, and it would probably be accepted as fact and used in an answer to a question.

I will admit though, that in the last paragraph, starting with “Why East Lane Stairs matter:”, ChatGPT did a much better job at summarising the reasons than I did – I need to work on using less words in my posts.

I will never be using AI in any posts, whether for research or writing, so they will probably continue as being too wordy and with some dubious grammar – but that is what makes us human, rather than an AI agent.