The US government is looking to use AI to help fight cybercrime

Another day in 2023 and another story about AI. This time, however, it isn’t about OpenAI or ChatGPT, this time it is about a new tactic the US Government is implementing in its battle against cybercrime.

In a joint effort between the Science and Technology Directorate, which works under the banner of the Department for Homeland Security and the Cybersecurity and Infrastructure Agency, the US government is working on developing an AI sandbox that will enable researchers to work together testing new and innovative ways of fighting cybercrime and detecting threats.

Part of this move is the Cybersecurity and Infrastructure Security Agency's Advanced Analytics Platform for Machine Learning (CAP-M), which will be used on-site and in cloud environments too. Both agencies put out a joint statement on the move saying:

“Fully realized, CAP-M will feature a multi-cloud environment and multiple data structures, a logical data warehouse to facilitate access across CISA data sets, and a production-like environment to enable realistic testing of vendor solutions […] While initially supporting cyber missions, this environment will be flexible and extensible to support data sets, tools, and collaboration for other infrastructure security missions.”

This means that the CAP-M will be used for experiments, with data analysis and correlation taking place in various ways in support of cybersecurity actions being taken across multiple organizations. The idea is that this will help them stay one step ahead of malicious actors who are constantly innovating and seeking new ways to break through their defenses.

A report in The Register claims that the security of the platform itself as well as all the related privacy issues that could cause problems are being taken into account. However, beyond that, information is thin on the ground about this will be done.

AI and machine learning will be involved in the project in a couple of ways. Firstly, as is often the case with these types of issues, the huge amounts of data that will need to be processed have led to the managers behind the projects deciding that a machine learning loop will be needed to automate certain workflows such as exporting and tuning the data. Also, however, the project will also involve testing various AI and machine-learning techniques to see how effective they can be.

While there are concerns being raised such as the model not being open or general enough to miss some types of threats, the project is being well-received at the Federal level. As the cybersecurity implications, the model of cooperation between the agencies is also being touted as something that could be repeated elsewhere in the past.

Advertisement