Over the past year, 62 per cent of organisations worldwide have experienced deepfake attacks involving social engineering or the exploitation of automated processes, according to a report released Wednesday by Gartner, a leading business and technology insights company.

Additionally, 32 per cent of companies reported attacks targeting AI applications through the manipulation of prompts.

The survey, conducted between March and May 2025 among 302 cybersecurity leaders across North America, EMEA, and Asia-Pacific, also revealed that 29 per cent of respondents faced attacks on enterprise Generative AI (GenAI) application infrastructure in the last 12 months.

Many organisations found chatbot assistants vulnerable to adversarial prompting techniques, where attackers craft inputs to trick large language models (LLMs) or multimodal AI systems into generating biased or malicious outputs.

“As adoption accelerates, attacks leveraging GenAI for phishing, deepfakes, and social engineering have become mainstream, while other threats, such as attacks on GenAI application infrastructure and prompt-based manipulations, are emerging and gaining traction,” said Prashast Gupta, director analyst at Gartner.

Despite growing AI-related risks, 67 per cent of cybersecurity leaders believe these emerging threats require significant changes to existing security frameworks. However, the report advises a balanced approach rather than sweeping overhauls.

“Rather than making sweeping changes or isolated investments, organisations should strengthen core controls and implement targeted measures for each new risk category,” Gupta added.

In a related insight, a report from Bain & Company estimates that by 2030, around $2 trillion annually will be necessary to fund the computing power needed to meet global AI demand. Despite cost savings from AI, there remains an $800 billion shortfall in funding to keep up with projected requirements.