Narrative attacks and deepfakes, what to know about top cyber-attacks trends

Forrester has said the likes of narrative attacks and deepfakes make the top five list of cybersecurity threats organisations can expect to face this year.

The trends follow as key security leaders believe 78 percent of sensitive data has already been breached.

The global market research company revealed the findings in one of its latest reports, Top Cybersecurity Threats, 2024.  

Narrative attacks, deepfakes, AI responses, AI software supply chain and nation-state espionage topped chief trends found by Forrester for this year.

Report lead author Forrester principal analyst Brian Wrozek said, “Cyberattacks have become so common that when an outage or disruption occurs, many people initially jump to the conclusion that a cyberattack must have been the root cause.

“It is easy to see why as the cybersecurity threat landscape has become a volatile mix of threats driven by rising uncertainty and increased complexity.

“Uncertainty generated by narrative attacks, deepfakes, and AI responses make it difficult to separate fact from fiction.

Wrozek added the AI software supply chain and “nation state espionage risks” create extra complexity to an “already convoluted threat environment.’

Wrozek also told Digital Nation earlier findings from a separate report, Lessons from the World’s Biggest Breaches 2023, noted sectors that are most targeted are public sector, education, and healthcare.

These sectors accounted for over 40 percent of the top 35 breaches we analysed from 2023, Wrozek added.  

“Financial services and insurance follows. Threat actors are often opportunistic so any company with security vulnerabilities and gaps in security controls will be targeted.”

Australian businesses can protect themselves through various methods, Wrozek told Digital Naton.

For narrative attacks Wrozek said minimising the impact “requires early warning”, done through “expanding your threat intelligence program to include monitoring open and dark web sources.”

“Countering deepfakes requires “using algorithms that detect manipulation of images; controlling the source of media and wrapping facial and voice biometrics with additional verification and protection layers”

This includes “behavioural biometrics, device ID fingerprinting/reputation, bot management and detection, digital fraud management and passwordless authentication.”

 Wrozek explained the trend of AI responsescan be tackled via “policy decision and policy enforcement points reside at the prompt level for applications with generative AI.”

For example, if using Microsoft Copilot or AzureAI, you may need to ensure you have the correct licenses to obtain logs with information on user prompts and what was returned.

New technologies have emerged to perform content analysis and filtering at this layer, like PrivateAI, Prompt Security, ProtectAI, and from data leakage prevention (DLP) vendors attempting to rebrand or pivot to offering controls in this space.”

AI software supply chain attacks could be solved organisations bringing “together representatives from security, IT, architecture, AI/data science, development, and the business to understand current AI usage and future AI integration plans and identify existing gaps.”

On nation-state espionage, Wrozek recommends consulting “the NIST National Cybersecurity Center of Excellence’s resources on cybersecurity for the space domain.”

“Lastly, use the Forrester Model To Defend Against Nation-State Threats to develop organisation-specific threat models and mitigation plans.

 “Regardless of your organisation’s role in contributing to or operating space systems, leverage NASA’s Space Security Best Practices Guide to establish end-to-end security control across space and ground assets,” Wrozek said.