Elon Musk: Artificial intelligence could become smart enough to take down the internet

Is the worst yet to come on cyber hacking? It’s a question that has been debated by scientists and technology experts around the world with some saying that artificial intelligence (AI) could be a threat to not only future technology, but also human existence.

And Elon Musk, founder of Tesla and SpaceX, is one of them. Musk has raised concerns about AI before, and his latest Twitter exchange shows he hasn’t changed his mind.

 Ironically, Tesla is using AI technology to build and test driverless cars.

advertisement

Musk’s tweet references an Economist article that speaks about the rising tide of DDoS attacks, why we’ve been so bad at stopping them and what we need to do to prevent it.

The entrepreneur’s Twitter exchange suggests he believes it’s only a matter of time before man-made AI becomes smart enough to learn how to take down the web.

Musk believes the internet is particularly vulnerable to attacks, especially if hackers are able to use “gradient descent algorithms” to bring it down.

Elon Musk (Marcio Jose Sanchez/AP)

A gradient descent algorithm is a process that finds the optimal solution to a complex mathematical function. Optimisation is the key factor in AI learning and computing – almost every machine learning algorithm has an optimisation algorithm at its heart.

However, the same process can be used by advanced AI to carry out a distributed denial-of-service (DDoS) attack – flooding websites with traffic until it can no longer support it and is forced offline.

And that’s not all. If multiple parties are using advanced AI in a cyber battlefield, Musk believes it could lead to an AI versus AI showdown.