Taiwan to share Chinese hacking attempts with private firms to train AI defences

Taiwan is looking to share information of hacking attempts by China with private companies to help train artificial intelligence systems to predict and prevent future attacks.

According to a report by the Financial Times, IT security companies will be able to use the data to train machine learning systems. The idea being that such data will help AI figure out when an attack is likely and put the necessary defences in place.

Jyan Hong-wei, director of Taiwan's department of cyber security, told the newspaper that companies would be able to get their hands on data going back years relating to attacks on Taiwanese government computers from nation-state hackers in the mainland.

Jyan said that the value of the data was "immense", and this would help AI spot patterns and "develop better tools". He added that currently, most cyber attacks relied on humans to code the appropriate hacks but soon software would undertake the job of coding attacks making them "more capable and much faster".

The report quotes an anonymous Microsoft executive saying that technology companies have been trying to persuade Taiwan to open up its database of cyberattacks for some time, and that this information would be "extremely valuable for our predictive threat analysis".

Jyan said that China often tests malware and hacking tools on Taiwan before using them on other countries. He added that the government in Taiwan will make an initial decision on what data to share and how to filter it to reduce security risks to itself.

"We will probably end up deciding that certain information can be widely shared, while other pieces can only be passed on to a certain party mandated by us on a contractual basis and which signs a non-disclosure agreement," he told the FT.