How would a robot fix the housing crisis?

Posted: February 16, 2023 | Author: julesbirch | Filed under: Affordable housing | Tags: ChatGPT, Housing crisis |Leave a comment

Originally written as a column for Inside Housing.

Question: Is there a housing crisis in the UK?

‘Yes, there is a housing crisis in the UK. The housing crisis in the UK is characterized by a shortage of affordable housing, high house prices, and a lack of social housing. This has led to an increase in homelessness and people being unable to afford to buy or rent a home. The crisis is driven by a number of factors including population growth, lack of new home building, and government policies that have favored homeownership over renting.’

As you may have guessed, it wasn’t me that wrote the last paragraph. The clues are, I hope, there in the plodding prose style and the American spellings.

The author was actually ChatGPT, the artificial intelligence chatbot that has workers anxious for their jobs and schools and universities fearing that essays will be written by robots rather than students in future.

The bot goes on: ‘The crisis has been exacerbated by the COVID-19 pandemic, which has led to job losses and economic uncertainty, making it harder for people to afford housing. The situation is particularly dire in London and other major cities, where housing is in high demand and prices are particularly high. The crisis is not only affecting low-income households, but also the middle-class families and young people. This is making it harder for them to access decent and stable housing.’

If that feels a bit out of date (the pandemic has arguably contributed to the opposite problem of labour shortages), it seems a reasonable summary if you accept that longstanding structural issues that we have done next to nothing to address can reasonably be described as a ‘crisis’. ChatGPT does at least get that the problem goes across incomes.

But what would the robot do about this, I wondered.

I asked ‘how can we fix the housing crisis in the UK?’ and got this answer:

‘There are several ways that the UK government and private sector could work together to address the housing crisis, including:

1. Building more affordable homes: This could involve increasing funding for social housing and incentivizing private developers to build more affordable homes.
2. Rent control: The government could introduce measures to control rents and make them more affordable for renters.
3. Tax reform: The government could reform taxes to make it more expensive for investors to buy properties and leave them empty, and more attractive for owner-occupiers.
4. Planning reform: The government could make changes to the planning system to make it easier for developers to build new homes.
5. Increase in wages: Increase in wages would help people afford housing.
6. Increase in supply: Increase in the number of houses available to buy or rent would help to lower prices.

‘Ultimately, addressing the housing crisis in the UK will require a multifaceted approach that involves both the government and private sector. It will also require cooperation between different levels of government, such as local and national government, to ensure that policies are implemented effectively.’

That seems a reasonable, if basic, summary that draws on most of the familiar fixes – and ChatGPT even goes where politicians fear to tread on tax reform.

The robot seems to get that we need many different solutions but there is no real sense that it understands the problem: increasing wages might make homes more ‘affordable’ but not for very long without greater supply and possibly not at all if accompanied by rising interest rates.  

That became even clearer when I asked it ‘why can’t I afford a home?’ Here’s the response:

‘There are several reasons why you may not be able to afford a home, including:

1. High house prices: If house prices in your area are high, it may be difficult to save enough money for a down payment and to make mortgage payments.
2. Low income: If you have a low income, it may be difficult to afford a home, even if prices are relatively low.
3. Lack of affordable housing: If there is a shortage of affordable housing in your area, it may be difficult to find a home that you can afford.
4. High debt: If you have high levels of debt, it may be difficult to qualify for a mortgage.
5. Lack of savings: If you have not been able to save enough money for a down payment, it may be difficult to buy a home.
6. Limited access to credit: If you have a low credit score or lack of credit history, it may be difficult to get a mortgage.
7. High unemployment or low wages: If you are unemployed or have a low paying job, it may be difficult to afford a home.
8. Lack of affordable rentals: If you are renting, and the rental market is expensive, it may be difficult to save enough money to buy a home.

‘It’s important to note that the factors which contribute to affordability of housing vary depending on different regions and demographic groups. It’s important to consider the specific reasons that make it difficult for you to afford a home. Additionally, there are various government programs that can help you to afford a home, such as Help to Buy and Shared Ownership, which you can look into.’

The robot assumes that this is all about personal finance but if Help to Buy and shared ownership are the answers then maybe there is a problem with the question?

It’s one I chose deliberately as an echo of the 2018 book by Josh Ryan-Collins, Why Can’t You Afford a Home?, that ChatGPT should add to its reading list.

There is no sense that the housing question might just be more complex than the bot makes out – and that the solutions might not be so simple either.

The politics of achieving those solutions is also completely missing, which comes back to my point about ‘crisis’. Many people, and a majority of those most likely to vote, are doing quite nicely out of high house prices and politicians are terrified of annoying them. See the government’s surrender to its backbenchers on planning reform for just the latest example of this.

Perhaps that’s not ChatGPT’s fault, though. The chatbot may be able to synthesise lots of answers from those that are already out there but expecting anything more original is probably asking too much.

Turning that point on its head, it seems that a chatbot can come up with a reasonable if limited summary of solutions familiar from countless reports on housing. Is it just ChatGPT that’s missing something – or is it us too?

---