The rapid evolution of artificial intelligence has introduced unprecedented speed and scale to cyber threats, leaving traditional defense mechanisms straining to keep pace. Cybersecurity leaders convening at the Cyber Security conference at Tel Aviv University addressed the urgent reality of high-speed AI-driven attacks, emphasizing that organizations have a matter of minutes to identify threats before facing catastrophic damage.

Industry experts outlined the narrow operational windows required to mitigate advanced breaches, alongside broader industry debates over whether to accelerate or pause AI development in the face of escalating vulnerabilities.

The Accelerated Four-Stage Cyber Defense Timeline

To mitigate catastrophic AI-based hacking against businesses or government agencies, Trellix Chief Technology Officer Mo Cashman stated on Tuesday that cyber defenders must identify such attacks within 15 minutes.

Cashman outlined an accelerated four-stage timeline designed to prevent the worst effects of rapid AI cyber threats:

  • Observe / Threat Identification: Must take place within 15 minutes to spot initial abnormalities or threats.
  • Orientation: Takes place within two hours to better understand the nature and implications of the threat.
  • Decision: Occurs within four hours to determine the necessary response and mitigation actions.
  • Action: Implemented within four to 72 hours, depending on the nature of the threat, to subdue and end the attack's effects.

Highlighting the urgency, Trellix pointed to recent PHP programming vulnerabilities disclosed in late October that briefly left an estimated 70% of the internet vulnerable until patched. Cashman noted that his team has occasionally managed partial remote remediation within 20 minutes by having pre-established digital security assets ready for a client.

The Expanding Cost and Safety Imbalance

The financial and strategic imbalance between attackers and defenders has widened significantly due to AI. AISLE Chief Information Security Officer and Chief Operating Officer Jaya Baloo told the conference that while attackers have always held an investment advantage, AI has worsened the disparity.

According to Baloo, perpetrators can launch semi-supervised attacks for as little as $25, while defenders must invest heavily to counter them. She attributed much of this imbalance to US and Chinese companies rushing out new models without accounting for potential dangers, arguing that governments must rapidly regulate and criminalize malicious conduct perpetrated by certain AI systems.

Conversely, CrowdStrike Chief Business Officer Daniel Bernard argued that the US, Israel, and the West must press onward with AI advancement without slowing down, despite security concerns. Noting that adversaries will not sign pause agreements, Bernard pointed to his company's development of an AI "Kill Switch" designed to manage gateways, cloud agents, and user permissions to restore safety without isolating development in a secure sandbox.

Adding a technical perspective, Microsoft Israel Director of R&D Michal Braverman-Blumenstyk emphasized that defenders must constantly study the internal processes of the "AI brain" rather than merely observing outcomes after a system goes rogue.

Military Applications and Future Outlook

Beyond corporate and government cybersecurity, military organizations are actively integrating AI to process vast operational metrics. Head of the IDF J6 & Cyber Defense Directorate Maj.-Gen. Aviad Dagan shared that the military has harnessed AI to analyze vast amounts of wartime data accumulated over three years of conflict.

Dagan noted that the military is moving toward using AI agents to help initiate and propose future strategies across complex terrains, though ultimate decision-making authority will remain firmly in the hands of human soldiers.