AI is already making online swindles easier. It could get much worse.

Here are six key points summarizing the article:

PromptLock Discovery: Researchers Cherepanov and Strýček unveiled PromptLock, an AI-powered ransomware they believed demonstrated the potential for highly flexible malware attacks.
Clarification from NYU Team: The following day, a team from New York University claimed the ransomware was merely a research project, designed to demonstrate the automation of a ransomware campaign rather than an actual attack in the wild.
Real World Threats: The discussion highlighted ongoing issues, emphasizing that while PromptLock was a research effort, real cybercriminals are increasingly using AI to lower the barriers to launching attacks.
Imminent Risks: Lorenzo Cavallaro, a computer science professor, stated that it’s “a sheer reality” that cyberattacks will now become more common and effective due to AI tools.
AI Misuse Beyond Malware: Experts suggest focusing on immediate risks like deepfake technologies and AI speeding up scam efforts, highlighting more current and pressing threats than purely AI-generated ransomware.
Scam and Spam Explosion: Since the rise of tools like ChatGPT, generative AI has been heavily utilized to create spam and sophisticated targeted email attacks, increasing from 7.6% to 14% in the first half of 2025 according to research from Columbia University, the University of Chicago, and Barracuda Networks.