GoKawiilThe key to security program effectiveness is optimizing remediation. This has become increasingly difficult as organizations strive to modernize their processes with innovative technologies, including artificial intelligence (AI). As employees gain capabilities to collaborate and work faster, cyber assets and attack surfaces proliferate, making it difficult for security teams to take the needed actions to mitigate risk.
Now, as organizations look to leverage agentic AI in areas such as software development, instead of incrementally increasing productivity, we are expecting exponential gains in productivity, further proliferating attack surfaces. At the same time, the threat landscape will also evolve rapidly, with attackers taking advantage of AI to scale their attacks.
Security teams need to keep the AI advantage on the defender side to win the AI arms race. The good news is that cybersecurity and asset context are accessible programmatically, allowing continuous API-based data collection. With AI-backed data analysis and agentic AI capabilities that can autonomously perform tasks, we are entering an exciting era of risk remediation in the age of AI.
Related:Xygeni GitHub Action Compromised Via Tag Poison
In fact, last week, we saw the introduction of Claude Code Security from Anthropic — using context and traces of data flows across files to identify vulnerabilities and aid in remediation. This offers many advances compared to traditional vulnerability scanning approaches, because the contextual data can help find security vulnerabilities that traditional scanning tools might miss. It was a smart move for Anthropic to release a security tool to support developers using Claude Code to build their applications. But how ready are security teams to embrace agentic AI?
Omdia’s recent study on "Automating Risk Reduction in the AI Era" showed organizations are rapidly moving toward AI-driven auto remediation. Most organizations (88%) are currently using AI-driven remediation, including 44% who said they have implemented AI-driven automated remediation for a majority of exposure types and 44% who said they have deployed AI-driven automated remediation for some exposure types and are actively exploring additional deployment.
The top types of fully automated remediation actions are:
Cloud infrastructure configuration changes (53%)
Network access controls (50%)
Identity and account permission changes (50%)
... continue reading