Reliably detecting humans on the Internet

1/ There’s ONE big unsolved problem for anyone who’s interested: detecting humans on the Internet.

It has potential to unlock billions of dollars of value. Here’s how.

2/ Today, some of the biggest tech companies (Google, Facebook, Twitter) make billions of dollars by monetising human attention.

Digital ad industry is total of $280Bn.

3/ The reason Facebook, Google make this amount of money is because their algorithms for detecting bots are closed source.

There’s a continual arms race between click-farms / bot-makers and bot detecting algorithms at these companies.

4/ Even then the problem of fake clicks in digital is massive. P&G recently pulled out of FB

Many other companies are following the suit.

5/ In the offline world, the cost of viewing an advertisement is high. To claim higher TV advertising rates, a TV channel will have to purchase millions of TVs, keep them ON and then somehow convince advertisers that those are genuine watchers.

6/ In the digital world, since ad viewing is happening on the Internet: it’s hard to convince whether at the other end, there’s a person who watched an ad or a bot.

7/ This problem worsens significantly in the blockchain world. Imagine that your bot-detecting code is open source, it’ll be trivial to beat it and claim advertising money.

8/ Reliably detecting human attention seems one of the positives going for centralised, closed source systems like FB/Google.

9/ Users will not prefer to pay for content directly as they’ve been used to free content, so advertisers are an important party in our digital economy.

10/ I’m not sure how this can get solved. An obvious answer is captchas but they get broken constantly (and isn’t a great user interface).

11/ Anyone who is able to detect humans reliably on the Internet will create billions of dollars worth of value in our economy.

Any takers? Your opinion?

Have any comments? Email me.