AI’s rising data costs and new threats signal a radical shift in cybersecurity’s future

Artificial intelligence is rapidly becoming the most valuable — and most vulnerable — asset inside modern enterprises, said Vasily Dyagilev, Regional Director for GCC, Saudi Arabia and RCIS at Checkpoint, speaking at the Gulf News Cyber Forum 2025.
Dyagilev revealed new internal statistics showing that “93 per cent of recent projects that were sold to the boards failed on the cost ratio, and most of them failed because the cost of data which was used to train up to date has increased up to 167 per cent.” With AI models becoming outdated in as little as 18 months, organisations are now facing soaring retraining costs and mounting cyber-risk.
He stressed that AI should no longer be viewed as a cost centre but as a core revenue-generating asset — making it a prime target for attackers. As AI adoption accelerates across industries, Dyagilev warned that language-based threats and prompt-driven breaches will define the next wave of cyberattacks.
Highlighting real-world leakage risks, he said experiments have shown that “with certain prompts, banking applications suddenly started to give you all the information about another user,” underscoring the urgency of securing AI agents and training data.
Dyagilev concluded that protecting AI models — and the data that fuels them — will become the “new frontier” in cybersecurity worldwide.
Sign up for the Daily Briefing
Get the latest news and updates straight to your inbox
Network Links
GN StoreDownload our app
© Al Nisr Publishing LLC 2025. All rights reserved.