Microsoft Report: Attacks That Use 4.5 Times More Effective AI

JAKARTA - Microsoft has just released an annual Digital Defense Report 2025 (MDDR 2025) report, highlighting changes in cyber threat patterns around the world amid the increasing role of artificial intelligence (AI).

In this report, Microsoft said that identity-based attacks still dominate, where more than 97 percent of identity attacks, for example, stemmed from mass codeword guessing attempts.

Furthermore, this report also shows where ransomware has now evolved into data extortion. If previously the perpetrators only encrypted the system, now they are also stealing sensitive data to sell or use as a means of negotiation.

Public sectors such as hospitals, educational institutions, and local governments are the most vulnerable targets due to limited security resources.

Meanwhile, the tech giant discovered that malware such as Lumma Stealer is now a new entry point for cybercrime.

This type of malware is usually used to steal user information, including passwords, session tokens, or personal data, through malvertising campaigns and search results manipulation (SEO poisoning).

In line with AI developments, criminals have now also used AI to accelerate search for vulnerability and multiply automatic phishing scale. Where attacks that utilize AI have a 4.5 times higher success rate than traditional ones.

"Indonesia's fast digital economy growth needs to be balanced with strong security readiness and discipline," said Dharma Simorangkir, President Director of Microsoft Indonesia.

To that end, Microsoft recommends five main steps to strengthen cyber resilience: