Las Vegas is one of the most bustling U.S. cities and because of the large casino operations in the city and electronic transfers of millions of dollars every day, cyber security is of prime importance. The city’s information security team, which comprises of only three people, relies on artificial intelligence (AI) to protect it against ransomware and phishing.

“The things that keep me up most are ransomware and phishing. Some of the simplest attacks but the hardest to defend against.” Las Vegas chief information officer Michael Sherwood told TechCrunch Friday.

Read: Americans Leading Ransomware Target And Most Likely To Pay: Symantec

One cyber defense firm, Darktrace, helps Sherwood in his job. The company has been using AI for cyber security solutions ever since it was launched in 2013.   

The system, which protects Vegas, is called the Enterprise Immune System uses machine learning algorithms created by a team of mathematicians from the University of Cambridge. The company says that its machine learning system is based on the self-learning intelligence of the human immune system and learns patterns of cyberattacks fast and prepare the city’s defense better.

The company says its system is engineered to work across all network types — physical, virtualized, cloud and most importantly the Internet-of-things and associated system.

But, what Darktrace is aiming at is interesting — it wants to create a system in which no human intervention is required. AI will be a be-all and have-all system for cyber security. Surrendering a city’s cyber security to machines essentially might sound an intimidating idea, whose time is still to come, but Sherwood seems bullish on it.

“Do you go in all the way or do you not? I wouldn’t live without artificial intelligence. Humans make wrong decisions every day,” he said to TechCrunch.

Is AI the solution to the increasing cyberattacks every year? We don’t know as yet. But what we know currently is that the scale and the number of cyberattacks are increasing every year. According to anti-virus company Symantec’s Internet Security Report, 2017, one in every 131 emails contains a malicious link. Many of these emails have executable files, which can encrypt your system, after which you might be asked for a ransom for getting access restored. Americans are the leading target of ransomware and the ones most likely to pay up globally.

One of the leading examples of how cyberattacks have evolved is the hack of the Democratic National Congress Chairman, John Podesta’s emails. How it happened was way simpler than you might think. Podesta received a mail, which looked like an authentic Google mail asking him to change his password to make his account secure. Podesta actually ran the mail through his security team, which cleared it and therefore he put his password in the window linked in the mail. In reality, Google never asked him to change his password and that Window was in no way connected to Gmail. His password was never changed, but he did give away access to his email and we all know what happened next.

Read: Who Is John Podesta? 10 Facts About The Hillary Clinton WikiLeaks Source

There is no doubt that it was a human error that led to the cyberattack. While the effectiveness of AI in cyber security is not known yet, but it might be an alternative and one of the steps needed to massively ramp up government cyber defense.