The deepfake threat is significant, and the consequences can be devastating and require insurance. As a Reuters report highlights, the need for cyber insurance matters – but insurance will only cover organizations to a limited extent, actively combatting deepfakes is critical too.

As cybersecurity measures struggle to keep up with the rise of deepfakes there’s a growing probability of attack success. Executives should consider proactively integrating deepfake scenarios into their crisis management plans to ensure they’re able to respond to an attack and mitigate the consequences.

Australia just added to the global momentum to regulate harmful deepfakes, setting a legal precedent in the Asia-Pacific region. New legislation imposes a duty to respond quickly to reports of deepfake porn.

As AI steadily becomes core infrastructure, protecting it is as mission-critical as securing networks or databases. The NSA’s AI Security Centre validates this and is set to help ensure AI systems can be trusted and safeguarded against emerging threats.

At the recent 2023 DEF CON hacker convention the cybersecurity community stress-tested LLMs, taking a proactive approach to uncovering AI vulnerabilities (from jailbreaking LLMs to prompt injection attacks) before malicious actors do.

Turning harmful actions into a criminal offence is sometimes the only way to deter bad actors. The UK government now took concrete action to tackle a malicious use of deepfakes that technical measures alone haven’t solved.

When election regulators sound the alarm on deepfakes, it underscores the extent of misuse of AI. Clarity believes transparency in AI-generated political content isn’t just ethical – it’s table stakes for preserving public trust in the electoral process.

We know we need guardrails around AI use – but can companies voluntarily agree to apply controls to reduce the risk of malevolent AI use? The US Government appears to have made some progress in achieving that.

Catching malware can be like finding a needle in a haystack. This report from LANL shows how AI can shrink that haystack – a win for companies trying to guard against a growing cybersecurity threat.