White House Launches Million-Dollar Cyber ​​Contest to Enhance Government Infrastructure Security with AI

JAKARTA - The White House on Wednesday, August 9 announced that it has launched a multi-million dollar cyber contest to encourage the use of artificial intelligence (AI) to find and fix vulnerabilities in US government infrastructure, against hackers using the technology for nefarious purposes.

"Cybersecurity is a race between attack and defense," said Anne Neuberger, the US government's deputy national security advisor for cyber and emerging technologies.

"We know that bad actors are already using AI to accelerate vulnerability identification or create malicious software," he added in a statement to Reuters.

Many US organizations, from health groups to manufacturing companies and government agencies, have been the target of hacks in recent years, and officials have warned about future threats, particularly from foreign adversaries.

Neuberger's comments about AI mirror those of Canada's head of cyber security, Samy Khoury, last month. He said his agency had seen AI being used for everything from creating phishing emails and writing malicious computer code to spreading disinformation.

The two-year contest includes about $20 million in prizes and will be led by the Defense Advanced Research Projects Agency (DARPA) - the US government agency responsible for creating technology for national security - the White House said.

Alphabet's Google, Anthropic, Microsoft and OpenAI - US tech companies at the forefront of the AI ​​revolution - will make their systems available for this challenge, the government said.

This contest indicates a formal effort to address emerging threats and which experts are still trying to fully understand. In the past year, US companies have launched generative AI tools such as ChatGPT that allow users to create convincing videos, images, text and computer code. Chinese companies have also launched similar models to catch up.

Experts say such tools could make it easier to, for example, carry out mass hacking campaigns or create fake profiles on social media to spread false information and propaganda.

“Our goal with the DARPA AI challenge is to encourage the larger community of cyber defenders using participating AI models to race faster – using generative AI to strengthen our cyber defenses,” said Neuberger.

The Open Source Security Foundation (OpenSSF), a group of US experts working to improve the security of open source software, will be responsible for ensuring "winning software code is deployed promptly," the US government said.