Researchers Teach AI To Find ICO Scams, Before The Scams Happen

With the escalating popularity of cryptocurrencies, many people are doing almost anything to take advantage of the trend, either in a good way or a bad way.

While there are a lot of ICOs (initial coin offering) out there that promise to make investors rich, most of them were either scam or failures. And this can be a problem for cryptocurrency investing since there is no definitive guidelines to avoid being scammed.

This is where a Chinese startup Shannon AI wants to use machine learning algorithms to help people out.

Partnering with researchers from Stanford, University of California Santa Barbara, and the University of Michigan, the company revealed a white paper detailing how an AI can be designed to spot cryptocurrency scams.

"We introduce ICORATING, the first learning–based cryptocurrency rating system. We exploit natural-language processing techniques to analyze various aspects of 2,251 digital currencies to date, such as white paper content, founding teams, Github repositories, websites, etc. Supervised learning models are used to correlate the life span and the price change of cryptocurrencies with these features. For the best setting, the proposed system is able to identify scam ICO projects with 0.83 precision."

This strategy involves scanning, gathering and understanding various information, to draw an overall picture of a coin.

Machine-learning method to know which ICOs are good and which are bad can be an encouraging great news for everyone, except for the scammers. However, there is a problem:

Websites, white papers, founding teams, social media accounts, repositories and others can be made to look legitimate and professional, with a relative ease. And there is a fact that researching these can fool anyone.

The problem can be made worse by the cryptocurrency communities themselves.

When an ICO promises great things and many people are into it, the communities tend to spreading positive messages. It's almost impossible to know about the legitimacy of a coin by scouting information from those that are involved in it.

And also due to the nature of these scams, a red flag can be easily dismissed by cryptocurrency investors. They may see the alert as a paid attack of FUD, even when they came from reputable news sources.

So here, the AI works in a blackbox to come to its conclusion. A machine learning model that has less prior knowledge about the world, can work in contrast to human designed systems that require massive involvement of human experts, who inevitably introduce biases.

What this means, it's reasoning should not be affected by external factors and only based on the data it receives.

This could make it more trustworthy. According to Shannon.AI, the algorithms don’t do anything a person couldn’t do on their own, but it can do it faster and with better accuracy.

But still, it's success ratio is not a 100 percent perfect. So people should trade with care. Investing in cryptocurrencies can indeed be fruitful. But with their volatile nature, things can go bad in an instant.

Published: 
15/03/2018