Mark Cuban argues AI-enabled defence can answer the scale of AI-enabled attacks

His Reuters remarks use a household-pest analogy to describe repeated attempts; they are an opinion about security, not a measured protection result.

Mark Cuban argues that discussion of AI-enabled attacks should also account for the technology available to defenders. In Reuters-recorded remarks from Nashville on October 23, the entrepreneur said he was optimistic that security teams could use comparable AI capabilities to counter attackers.

He used a household-pest analogy to describe repeated attempts to find an opening: a large number of attacking agents could keep searching until one succeeds. His response was that defenders could deploy AI at scale as well.

An analogy about volume

The remarks express a view of the balance between attack and defence. They do not report a controlled experiment, prove that every AI attack can be stopped or establish the effectiveness of a particular security product.

The important distinction is between a system’s ability to generate many attempts and the outcome of those attempts. A large volume of activity can create work for a defender without necessarily producing an intrusion.

Defensive capability still needs an operating process

A security team has to decide which signals require attention and which actions an assistant may take. Automating a response can save time, but the response also needs to be appropriate for the affected service and reversible where possible.

Cuban’s optimism therefore raises a concrete question rather than settling it: can defenders use capable agents while keeping their own access and actions under control?

The Reuters transcript establishes what he said and when. Claims about actual security improvement require evidence from deployed systems and relevant evaluations, beyond the analogy in the interview.

Sources

- [Mark Cuban discusses AI attack and defence agents](https://uk.marketscreener.com/news/mark-cuban-says-ai-hack-agents-are-like-cockroaches-but-can-be-controlled-ce785cdfdd81f52c) — Reuters via MarketScreener