Data and context values switch the value of the model
Consider the following scenario: Financial services companies have built large and expensive AI models to analyze user behavior and recommend product changes and new features. The company also has a small, relatively simple AI model used to analyze the financial activities of large customers and create liquidity recommendations. Which models need to receive security priorities?
Given this context, the obvious answer is the smaller AI model, as financial services companies want to protect their clients from the trends of dangerous business email compromises and spear phishing attacks. To this end, security investments must be scaled by data values rather than model features, as attackers perform cost-benefit analysis before attempting a violation. Processing financial data that is sensitive to relatively simple AI models requires more robust cybersecurity and protection than cutting-edge AI models that generate creative content.
Rather than targeting the most sophisticated AI systems, cyberattackers consistently choose the least path of resistance to valuable outcomes. This value-based approach to security investments aligns resources with real business risks rather than technical complexity. Organizations can optimize cybersecurity and resilience spending by protecting their most valuable assets, regardless of the refinement of their underlying AI systems.
