Wonga is a great example of a company taking advantage of people who are too desperate to have a choice and feel that their local loan-shark is too inexpensive for them.

The banks are not scared, they're just currently unwilling to plumb the moral depths that Wonga is.

Wonga is a rather poor example. They charge interest rates > 4000%. At that APR you can hand out loans to fourth graders and still make bank.

edit: first attempt at finding a payday lender with cheaper rates was a success: 600% http://www.paydayone.com/texas-loan-cost-and-terms.aspx

Perhaps, but the banks are also big Big Data users. My old company, Causata, had lots of relationships with banks wanting to unlock the value in their data.

It's true that startups will always challenge big businesses (such as banks), but don't think the banks aren't data savvy. In fact, they have a tremendous history of using computer systems to process and analyze data.

I can't read this kind of thing without wondering who these ML experts are that feel personally fulfilled putting their skills to work for 4000% APR payday loans. ML skills are in demand in a lot of industries that aren't directly or openly fucking the desperate.

The payday loans are only that high because the associated risks are immense. And this is precisely why you need to have ML experts working on this problem.

First - I don't know anything about Wonga, nor do I think that vulnerable people being forced into needing a loan at enormously high interest rates is somehow a good thing for society.

However, as I read the comments here simply condemning payday loans without further explanation, I find myself wondering whether anyone has a better suggestion?

A 4000% interest loan might seem like a terrible deal, but simply being evicted, or not being repair a vehicle essential for work could far more costly. If happens occasionally and it really is paid back at payday, I can see it being a lifeline.

The article suggests that the interest rate is basically directly related to the default rate, and that ML is enabling better credit assessments, and thus lowering the default rate and interest rates correspondingly. The fact that there are now competitors in the market should keep pressure on this.

It sounds to me that the ML experts are contributing to making loans available more cheaply to those who need them and can pay them back, potentially helping people who have a job not to fall into even worse poverty.

Again, I think that the poverty trap is an abhorrent thing for society, but it sounds as though ML experts are making the payday loan industry less harmful. Am I missing something?

Ugh. Don't they have laws against usury in UK?

Disappointingly not. In fact recent government commissioned studies recommended not regulating this further for fear of driving customers to more unscrupulous lenders (presumably ones without flashy offices and good marketing - not actually more expensive ones, which are hard to find).

It's pretty disgraceful. Thankfully it is beginning to be flagged up as more of a problem as some of the better newspapers are beginning to do some proper investigative reporting.

How do Wonga et al tie a borrower to his facebook/twitter/osm account? And then how do they get the data for an account?

They buy it.