Agentic AI, a new form of autonomous artificial intelligence, is projected to reduce the time required for account takeovers by 50% within two years, according to Gartner. This technology automates steps like deepfake-driven social engineering and credential compromise, posing significant cybersecurity risks. The technology also enables more sophisticated and large-scale phishing attacks, particularly through email. These attacks build on the capabilities of generative AI, which has already enhanced phishing tools. Gartner predicts that deepfake audio and video will be used in 40% of social engineering attacks by 2028. Separately, a phishing campaign has shifted from targeting Windows users to Mac users, exploiting typos in URLs to redirect victims to malicious sites. The campaign, identified by LayerX Labs, uses the Windows.net infrastructure and targets Safari users, following a 90% drop in Windows attacks after new browser protections were introduced. Experts recommend measures such as phishing-resistant multi-factor authentication, AI-powered detection systems, and employee training to address these threats.