A Security business has issued a caution about the fall of highly personalized, AI-powered digital risks that could substantially increase scams, phishing, and influencing procedures in 2025 and beyond.
” As relational AI becomes more integrated into corporations and society, we may stay vigilant. Hyper-personalized attacks and AI usurpation may necessitate industry-wide cooperation to overcome them. Business leaders should know that computer risks are no longer independent concerns, they are business risks with far-reaching proper implications”, said Ian Felipe, state manager at Trend Micro Philippines.
In the 2025 predictions report from Trend Micro, the development of malignant “digital twins” was emphasized. Bruted or leaked personal information is used to teach large language models ( LLMs) to imitate people’s knowledge, personality, and writing styles. These AI-powered tools, combined with damaged biometric data and deepfake video/audio, could lead to targeted social engineering attacks and identity theft.
Deepfake technology and AI will help large-scale, hyper-personalized cyberattacks, including increased business email compromise and “fake employee” scams, targeted fiscal scams, including “pig butchering” frauds, AI-driven romance scams that manipulate victims before passing them to human operators, improved open-source intelligence gathering for adversaries, and large-scale disinformation campaigns using fake social media personas.
Additionally, as more and more companies adopt AI, they must prepare for new security issues, such as an AI agent being hijacked to hack or spoof enterprise systems, as well as unexpected data leaks from relational Artificial tools and Iot resource consumption that causes denial-of-service attacks.
Memory problem flaws, API flaws, container escapes, and XSS and SQL injections are different pressing cybersecurity threats. Cryptocriminals will develop strategies to defy node monitoring and response tools.
Cyber threats
To overcome these escalating risks, Trend Micro recommends businesses take a strategic, risk-based method to cybersecurity. Additionally, it recommends using AI for risk intellect, asset management, and attack path prediction, as well as centralized danger identification and mitigation strategies.
The company even advised businesses to improve employee training to recognize and respond to AI-powered digital threats and protect AI technologies from misuse through thorough verification and multi-layered defenses.
Additionally, it advised the company to improve LLM security by implementing sandboxing, tight data verification and protection against fast injection attacks, ensure supply chain security and implement multi-layered network defenses, and maintain visibility into AI-driven security risks and attack paths. / KOC