Do We Need To Use AI To Fight AI?

Artificial intelligence may be the only way to defend against artificial intelligence


Necessity is the mother of invention. We know this from the huge number of innovations that have solved problems that consumers have, whether that’s Uber solving the problem of hailing down a taxi, through to the iPhone solving the problem of having to lug around multiple devices to each do a different job. Necessity has seen some of the world’s most powerful technologies being created to compete with other technologies. For instance, one of the key reasons why data-driven projects have become increasingly powerful is because competitors keep finding more powerful technologies, so their competitors then need to find a way to do things even better, creating a cycle of innovation and improvement.

The same thing is currently happening with artificial intelligence, where organizations are constantly doing battle to out-innovate one another in the quickly growing space.

One of the areas where this is taking place is in the sneaker buying market, outlined in Forbes’ article ‘How Bots Bested the $1 Billion Sneaker Resale Industry’ ( The US footwear industry generated $17.5 billion in 2016 and the resale market alone is worth over $1 billion. This is little surprise given that the Slim Shady’s Js, released in 2015, in a limited run of 10 pairs each sold for between $18,700 and $30,100. Even some of the best selling shoes of all time are sold for thousands of dollars, with the 1985 Air Jordans, which sold for $65 when released, now selling for close to $2,800.

There is big money in reselling sneakers, which means that when companies like Nike, Adidas, and Reebok release limited editions, those who want to profit from them through resales cannot risk missing out. As there is intense competition for these shoes (they often sell out in under 1 minute) resellers have turned to AI driven bots to increase their chances of beating the queues.

This is clearly something that companies do not want to happen, after all, their customers will become increasingly frustrated if they consistently need to pay well above their original asking price to get the shoes they want from resellers. So large companies like Nike constantly update their sales method to try and identify when a bot is trying to buy and when it’s a human. This is why you see the increasingly common ‘prove you’re not a robot’ tests before you buy something.

An attempt by bot makers to get around these barriers is counteracted by companies beefing up their security to identify them, which then creates a cyclical battle with each side increasing their capabilities to fight the attempts of the other. AI is key to this fight as bots can now identify logical tests, identify the ‘tests’ and over time and billions of attempts (according to Imperva, close to 52% of web traffic came from bots in 2016) find ways to solve them.

It isn’t simply shoe companies that are finding this issue either, but it is similar for any product with a limited run with a strong consumer interest. The ticket resale market, for instance, is worth around $5 billion, and tickets for the most popular shows are purchased through the use of bots as soon as they are released and often listed minutes later on resale sites with markups often well above 200% for the most popular tickets. When Thom Yorke, lead singer of Radiohead, found out that tickets for the band’s 2016 Madison Square Garden show had sold out in under 30 seconds mainly due to bots and then appeared on reselling sites for 300% more only minutes later, responded after frustrated fans complained to say ‘I'm as f---ed off as you are.’

AI is key to both this practice and it’s policing. Whether you agree with the idea of resellers being entrepreneurial or companies selling only to dedicated fans, one element that is winning out is the development of AI for both. 

Looking small

Read next:

Expert Insight: 'An Effective Visualization Results From A Great Deal Of Curiosity And Exploration'