Artificial Intelligence (AI) in London

I am also very concerned:

• that generative AI is contributing to a culture of misinformation in a uniquely troubling way, given the one-on-one nature of eg Chat GPT/Claude, but also the lack of accountability for AI agents elsewhere (companies repeatedly saying ‘oh dear, sorry, that was our AI- we’ll review our guidelines for it’ as if that means anything), the spreading of misinformation via generated art even by usually trusted sources (e.g. like educational history platforms like Dan Snow’s History Hit show using misleading AI art of historical figures despite having contemporary public domain photos and art they could have used). I won’t elaborate further here because hopefully you already have a wealth of information about this. This is also just a very depressing comment to write, apologies.
• about the now massively accelerated growth of technical debt, which is already a huge, very expensive vulnerability in our city’s, our national and our global infrastructure.
• About what is advertised and widely accepted about AI productivity, and what is being shown by evidence in multiple areas. (also related to cognitive debt, cognitive surrender). In July 2025, METR, a nonprofit research organisation focused on evaluating AI capabilities, published the most rigorous independent study of AI coding productivity conducted to date. Before the study, developers forecast that AI would reduce their task completion time by 24%. After completing the study, they still estimated that AI had made them 20% faster. But the developers using AI tools actually took 19% longer to complete tasks than those working without them. A 43 percentage point gap between subjective experience and objective performance. (Joel Becker et al. (METR), "Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity", arXiv, July 2025)

AI is here to stay. London needs to take advantage of it while it's still affordable.

Grants for entrepreneurs:
Stimulate the environment for people with great business ideas. Especially in fields that benefit the city. It can be anything from apps to help those with disabilities, to startups in finance, data analytics etc.These kinds of companies will hire a lot more along with providing services that help out Londoners.

Training for software engineers:
AI has unlocked unparalleled levels of software engineering speed. Put together a team to train up whoever wants to learn. And have them join teams that are in sore need to software engineers. We're talking NHS, GOV.UK, GDS, HMRC, the list goes on. Every additional engineer can create orders of magnitude more value. Stop using private companies and paying exorbitant levels of public funds to them when in-house engineers can do the same thing with a little training and help. Have a program where people get internships with the government after this training program. We do not need private companies to help us with AI. It's not that hard to use.

Embrace AI for Law making:

Even forums like this can harness AI. We can train models to sift through thousands of responses. We can make sure it weighs responses heavier for people who have more credibility in the field. We can detect when people are acting out of malice. This can be done for any public forums. It is so much easier to understand what the public actually want.

Public courses on AI misinformation:

It is very important the children and adults alike understand how to critically think about misinformation. This doesn't just apply to AI, but is now unfortunately more important than ever as it is even easier to create malicious content using AI. Enforce strict policies on all entities (especially social media companies) on labelling AI generated content clearly, and inform users when they may be looking at AI generated content. People cannot simply believe everything. 

Re: AI in my current life as a Londoner, and in London’s future, I am very concerned:

• that studies are already showing cognitive debt and cognitive surrender- i see this when speaking with people, scarily.
• that generative AI is changing culture around creativity, problem solving. E.g. friends who use chatGPT regularly both for work and personally- when I bring up an idea or question, I’m immediately met with ‘just ask AI/Chat gpt’ - e.g. options for an anniversary present (details of a watercolour painting). They don’t want to shut down the chat- they just always see ‘ask AI’ as the immediate ‘go to’ (cross-generational- 20s-60s) + look surprised when i ask what THEY think first. Even for fun or thoughtful things. (This sounds like a bad preachy scifi book with a transparent allegory now I read it back, but it’s real life. It’s so alienating.)
• that AI is promoting loneliness and isolation, especially amongst young men. Generative AI is being used by companies specifically advertising ‘virtual companions’, but also mainstream generative AI can be ‘personalised’, in that it can be given a name, told to act more friendly/jokey/affirming etc. The dangers and consequences of this are already being shown and others much more eloquent than me are enumerating the risks of this specific technology, as well as others sharing how this is changing how people are acting.
• that companies are choosing blunt AI tools over investing in, training people (see: the coding industry), the devaluing of and underestimation of people’s creativity, ingenuity and problem solving skills, and that the consequences of this in future will be a vicious cycle where this lack of investment in people + deskilling will be used by the same companies (already starting to do so) to argue further the idea that AI is superior to human work, that people aren’t bringing enough to the table, will pay even lower wages + poorer QoLs

There are some major issues with rapid AI adoption that need to be taken seriously.

1. The business model is unstable.   Most AI providers are currently operating at a loss to build dependence and collect data. What happens once businesses and individuals rely on these systems for daily operations, and the providers decide to raise prices to unsustainable levels? At that point, we either pay whatever they demand or face disruption to essential work. This is not a healthy or resilient foundation for London’s economy.

2. The ethical concerns are significant.   Many widely used AI models were trained on data, writing, and images that creators never consented to share. Encouraging large‑scale commercial use of systems built on unlicensed material raises serious moral and legal questions that cannot be ignored.

3. The employment impact is already visible.   There is no realistic way to prevent companies from cutting staff once AI adoption becomes normalised or encouraged. The current job market already shows how quickly entry‑level roles are disappearing, partly due to AI. If London accelerates this trend, we will see further reductions in opportunities, especially for young people. Jobs are not just productivity tools for corporations — they are how people sustain their lives. Large‑scale displacement will inevitably create knock‑on effects in crime, social welfare demand, and community stability.

London needs strong legislation to regulate AI use before these risks become irreversible.  Without clear limits and protections, we risk building a future that is economically fragile, ethically compromised, and socially damaging.

I'm writing this as someone who is just beginning a PhD in Scientific Machine Learning for Climate & Weather Modelling. 

My overall take is that AI offers a great deal of opportunity, but with significant caveats. 

Where I am most hopeful is in the application of AI to scientific and deep tech domains, where we are often already burning through vast amounts of compute (and all the energy, water, etc. that entails) to run complex numerical models. An example from my own area of expertise would be the potential to reduce the computational overhead on high quality weather forecasts from something the size of the Met Office Supercomputer to a single GPU. The same is true for running simulations in molecular science, aerodynamics, nuclear fusion and more. 

Beyond those deep tech problems where we have been stuck at certain bottlenecks for years, I think AI is offering the potential for productivity increases, but I think a healthy dose of cynicism is required to separate true productivity increases from the illusory ones. 

Some areas where I have thoughts:

• Software Engineering - AI seems good for writing basic code, for engaging in dialogue about system design, for writing documentation, and solving simple bugs, but falls down on large projects, can produce code that is overcomplicated, and through poor management can lead to significant tech debt once enough code is introduced without having anyone to explain how and why it was written the way it was.
• Creative Industries - Largely negative applications that seem determined to reduce the value of artists, though there are some exceptions (e.g. tools that help video editors spot differences between frames)
• Policing and Public Safety - All quite dystopian, especially when it comes to things like automated facial recognition, number plate reading, etc. 
• Access to public services - I'm hopeful here. It can be very complex to understand and access what rules exist in each borough, services, who to contact etc. and AI tools could help

I'm worried about overreliance on AI. It seems like people are unable to do anything without asking AI first. I sometimes feel unsure about doing anything without double checking with AI if it is the correct approach. If this continues, we may not be able to draft basic emails on our own or even write simple paragraphs. A lot of people also use AI for mental health and therefore are more reliant on AI which is unhealthy.

AI has its positives and negatives but it is the new way forward in most public services. The main worry for me is that it will affect peoples' jobs and that it may not be as accurate as we want it to be.. 

AI is an exploitative tool in its current state, is known to be harmful and riddled with misinformation, and incorporates works that have been taken without the knowledge or consent of the original creators. It is entirely unethical in its current state, and furthermore has almost irreversible climate and ecological impacts. There are many sources showing the effect of AI data centers on the environment and the people and animals living near to them. Furthermore, if an AI centre were to experience some sort of explosion, i worry that it may one day be known as a tragedy the likes of Chernobyl. AI is costing people their jobs, and in some cases even their lives, and in this market economy wherein inflation is rising exponentially and minimum wage is barely creeping up, it will lead to mass unhousing, poverty, etc even more extreme than what we are already facing. As citizens of London themselves, i am certain that the government is aware of the recent heatwaves and unpredictable weather events in this city, and even the country beyond that. They cannot pretend to be unaware of the rise since the implementation of AI on a global scale. 

I think my points are very clear. We do not want AI implementation, especially unregulated and in its current form. It is harmful in every facet!

Incorporating generative AI into the workplace will eventually push thousands out of jobs as they find themselves replaced by the very AI that their companies have told them to incorporate. Whilst it is cheaper for companies to use AI in many human roles, it is more environmentally wasteful and often much less efficient. It is not enough to not engage with generative AI in workplaces, but it must be actively pushed against - the “inevitability” of AI is nothing more than a self-fulfilling marketing tactic.

Generative AI is bad both socially and for the environment. It requires the building of huge data centres which are noisy, use large quantities of clean water and disrupt local wildlife and communities. The over-reliance on services like chatGPT and Claude causes widespread misinformation, danger to children online, and poses a solution to a problem that was already solved by search engines. 

So the thing about AI at the moment, is that it is like having a few thousand interns at your disposal. Great for the grunt work, but you would double-check all their work and never let them near anything important or confidential. AI makes things up and can offer out of date information as current. The key to using AI is know what the question should be, and how to ask it. Otherwise it is just garbage in, garbage out. What I really worry about is the lack of tax planning. The less work there is, the less revenue from income tax and national insurance. For the last 200+ years we have slowly been replacing unskilled workers with machines, and that genie isn't going back in the bottle. The definition of "unskilled" is now moving from farm labourers to junior lawyers. So we seriously need to consider where the tax money will be coming from in the next 200+ years, because clearly relying on income tax and national insurance is an assumption we might not be able to make for much longer. VAT is also doomed as a revenue source if we don't have any money to spend. Remember that VAT is only paid by people, not companies. There really aren't a lot of options left. Corporate tax rates could be raised, but those who currently avoid paying it will simply find new ways to avoid paying it. We need a new tax revenue source, and we should seriously consider taxing AI and robotics. 

The London Mayor has no realistic influence on the matter. Unfortunately politicians with access to public money do not feel any restraint in giving it away to wealthy corporations in the form of subsidies or controversial planning permission at the expense of the vast majority of Londoners, current and future.

We were doing fine without AI, weren't we? We were able to do our jobs properly, understand what we read, write emails, summarise using our brains, find information. "The need for AI" is an artificial creation by corporations set on making money by gathering data, stealing the creative and scientific work of millions of people, and displacing workers. All this without a care for the environmental impact.

I put AI to the test yesterday and asked it for the name of my husband. I was told that I was married to someone I have never met! How can one trust something that hoovers up any old information. I would love us to stop being bombarded by the AI industry. Who does it really benefit? 

AI is full of errors. 

I worry that young people will use it for job applications giving a homogeneous submission. Employers need to feel the personality of the future employee not an AI generated text.

AI is a relatively new tool and it has made us capable of taking informed decisions in minutes. I have found it extremely useful in answering me in areas which are unknown