Articles - Drop it like a bot


One of the most staggering statistics I’ve come across recently is that internet traffic is now only half human. Half (49.6%) of all traffic is bots . The full Imperva Bad Bot Report is fascinating and well worth a read. Given the recent trend, it seems likely that human traffic has since fallen below the 50% mark. The implications are fascinating and potentially quite dystopian. But as actuaries, we shouldn’t get carried away without doing some maths.

 By Alex White, Head of ALM Research at Redington

 

 Figure 1 Source: https://www.imperva.com/resources/resource-library/reports/2024-bad-bot-report/

 There is significant variation, especially around 2014-2015. The explanation for 2014 is changes to search engines, using more data scraping- while for 2015 it’s a huge influx of new users, largely from China, India, and Indonesia. That’s important, because it’s not a repeatable event - there’s not the same number of “fresh” humans who can start using the internet.

 All of this motivates some tests of statistical significance - is there a trend down or is it just noise? We looked at human traffic percentages and ran a simple t-test, which gave a p-value of 80%. Even starting in 2015 (to account for the 2015 surge of new users), the p-value is still 75%. So while it looks bad, there’s no significant evidence of a trend- it could all just be noise. So should we celebrate?

 Perhaps not yet. We also ran the same tests on bad bot traffic. Bad bots are, as the name suggests, bots designed with malicious intent; they cover a huge range of sins, from scalping, to credential stuffing, to outright fraud (‘good’ bots do things like indexing web pages).

 When we look here, again we see no statistical significance, with a p-value of 28%. However, starting in 2015, the p-value is a far more poignant 1%. Even with so few data points, the rise is statistically significant if the data is taken from anytime between 2016 and 2020, and even with 2 data points the p-value from 202 is still 7%. Bad bots are on the rise.

 This is perhaps more concerning in the context of how quickly the internet is growing. About 40% of the data in the world was generated in the last 2 years. As an aside, 90% is often quoted, but that seems to have been true around 2012, but not since, not least due to the enormous volume of videos, which make up around half of all data.

 
 Figure 2- Source: https://explodingtopics.com/blog/data-generated-per-day

 So we are producing more and more data, while traffic is more and more driven by malicious bots. It’s hard to see AI helping rather than hindering this problem too. From one perspective, we are making the internet a larger and larger Darwinian environment and are being outcompeted by bad bots.

 So what are the takeaways? When Willie Sutton was asked why he robbed banks, he gave the eminently sensible answer that “that’s where the money is”. In a similar vein. 37% of all ATO attacks target the financial services industry . So don’t scrimp on cyber security.
  

Back to Index


Similar News to this Story

Actuarial Post Magazine Awards Winners Edition December 2024
Welcome to the Actuarial Post Awards 2024 winner’s edition and we hope you enjoy reading about their responses on having won their award. The awards
Guide to setting expense reserves under the new Funding Code
The new defined benefit (DB) funding code of practice (new Funding Code) requires all schemes to achieve funding levels that ensure low dependency on
Smooth(ing) Operator
Private equity can be a great asset. It’s generally the most significant way to have any real world impact as an investor (eg infrastructure assets li

Site Search

Exact   Any  

Latest Actuarial Jobs

Actuarial Login

Email
Password
 Jobseeker    Client
Reminder Logon

APA Sponsors

Actuarial Jobs & News Feeds

Jobs RSS News RSS

WikiActuary

Be the first to contribute to our definitive actuarial reference forum. Built by actuaries for actuaries.