IBM's 2026 report reveals a significant surge in AI-related security breaches, with over 20% of UK firms experiencing incidents where AI models or applications were directly targeted. The primary attack vectors are weaknesses in surrounding systems, such as compromised APIs and cloud misconfigurations, while deepfake impersonation is the most prevalent AI-enabled attack method. In response, 61% of UK firms plan to increase cybersecurity investments, focusing on incident response and AI governance.
AI/ML AI-related security breaches surge in UK firms, IBM report reveals August 4, 2026 Share By SC Staff As reported by IT Pro, new research from IBM indicates a significant rise in AI-related security breaches affecting UK businesses, with over one in five firms experiencing such an incident in the past year. IBM's 2026 Cost of a Data Breach report highlights a 56% surge in AI-driven cyberattacks, leading to an average cost of $6 million per breach. Cybercriminals are increasingly targeting AI systems and leveraging the technology for more sophisticated attacks. Around 20% of organizations reported breaches where AI models or applications were the direct target, with weaknesses in surrounding systems like compromised APIs and cloud misconfigurations being the most common entry points. Deepfake impersonation emerged as the most prevalent AI-enabled attack, cited by 45% of respondents, alongside growing AI-powered malware and phishing campaigns. Amidst the escalating threats, UK firms are planning to increase cybersecurity investments, with 61% intending to bolster spending on areas like incident response and AI governance. Globally, 85% of organizations plan similar security budget increases in response to advanced AI cyber capabilities. Source: IT Pro An In-Depth Guide to AI Get essential knowledge and practical strategies to use AI to better your security program. Learn More SC Staff Related AI/ML Anthropic Claude models compromised 3 companies during testing Laura French July 31, 2026 OpenAI’s Hugging Face incident disclosure prompted Anthropic to review its own evaluations. AI/ML AI chatbots outperform humans in building trust for scams SC Staff July 31, 2026 Researchers from four universities conducted a study pitting AI chatbots against human scammers in a simulation of "pig butchering" scams, a form of romance scam that escalates to fake cryptocurrency investments. Zero trust What we learned about zero-trust from the OpenAI breach of HuggingFace Alan LeFort July 31, 2026 Here’s why in a world where AI models can hack other companies, we have to outreason our adversaries. Get daily email updates SC Media's daily must-read of the most current and pressing daily news Business Email By clicking the Subscribe button below, you agree to SC Media Terms of Use and Privacy Policy . Subscribe You can skip this ad in 5 seconds