Shin Bet removes social media accounts to block foreign interference in upcoming election

The offending accounts were deleted, and Shin Bet said it would continue to work to preserve the purity of the elections and democratic process.

Previously, and regarding previous elections, Shin Bet and other officials have named Russia, Iran, and China as top interfering parties.

The announcement came not long after a major warning on October 8 by Israel National Cyber Directorate (INCD) Chief Yossi Karadi to world leaders and heads of cyber defense organizations about the risk of losing control over AI agents and models, and his urging nations to urgently establish clear boundaries of sovereign authority.

“We are approaching a reality in which cyber defense systems will need to make decisions at a pace that does not allow human involvement in every action,” he writes.

Karadi’s warning follows recent incidents around the world in which AI agents breached government databases, alongside numerous other cases in which companies lost control over their agents’ actions.

What would happen if AI agents breached aspects of essential election functions or infrastructure indirectly needed for election day to take place, including transportation for voters to voting booths?

A summary statement said that Karadi called on national leaders and heads of cyber defense organizations worldwide to "determine now which powers may be delegated to AI systems and which decisions must remain exclusively in human hands."

"He notes that AI systems are advancing faster than nations can adapt their policy, regulatory and oversight mechanisms," said the statement.

The paper’s central warning concerns national security: as cyberattacks become faster, more automated and more targeted, "defense systems will also need to detect, decide and act at machine speed, beyond the point at which human approval can be obtained for every action."

Karadi writes that "the paradox, however, is that the more autonomy a system is given, the harder it becomes to understand its decisions, stop its actions in time and determine who is accountable when something goes wrong."

In the paper, Karadi presents a scenario that illustrates the risk: "an autonomous cyber defense system detects an attack exploiting a vulnerability for which no security patch is available, known as a “zero-day” vulnerability, and must isolate infrastructure within seconds."

A correct decision could stop the attack, but a false detection could unnecessarily shut down critical infrastructure, such as a power grid or a hospital.

To address this challenge, Karadi proposes that nations "adopt a strategic AI triangle that balances three interests: innovation, security and human control."

"Under this model, the question is not whether to adopt AI, but where to accelerate its use, where dependence on external providers and infrastructure creates a national risk, and which powers must never be transferred to machines," he writes.

Karadi proposes that nations "establish a clear hierarchy: actions that may be fully delegated to AI; actions permitted only within predefined limits; actions requiring human approval; and actions with irreversible national consequences that must never be delegated to an autonomous system, designated as “non-delegable sovereign authority."

“Leaders do not need to become AI engineers. They need to decide where dependence is acceptable, where resilience is non-negotiable, and where human authority must remain decisive,” Karadi writes. “Do not leave it to others to define your future. Enlist them to build the world you believe in.”

Netanyahu warns about potential election interference from Iran

“If they do, it’ll be a big mistake for them because we’ll respond with force and might they can’t even imagine,” he warned.