Deepfakes are mimicking executives. Voice clones are impersonating world leaders. Phishing emails now sound more authentic than the people they claim to be.
In the past year alone, 58% of organizations hit by a cyberattack suspect AI was used to enhance the intrusion. From deepfake scandals dominating headlines to the Qantas breach that exposed more than 5.7 million customer records, no public figure, industry, or organization has been immune to these evolving attack methods.
Each incident underscores a single, defining reality: artificial intelligence isn’t just influencing the threat landscape—it’s driving it. Attackers are faster, smarter, and harder to trace. 71% of cybersecurity leaders say attacks are growing more frequent, and 61% cite greater severity. Behind every breach that captures attention are dozens more that go unreported.
Only 24% of cybersecurity leaders feel very confident their teams can detect and respond to AI-driven attacks in real time—a gap that exposes how far defenders are falling behind. And with programs like CISA’s information-sharing initiative set to expire, leaders now face this new wave of threats with less visibility and fewer guardrails. If we’re going to regain confidence, we first need to understand how AI is reshaping the offense and why it’s giving attackers the upper hand.
Inside the AI Playbook: How Attackers Are Outpacing Defenders
Cybersecurity leaders say AI has changed not just the speed of attacks, but their nature. Routine threats have evolved into scalable, automated campaigns capable of overwhelming defenses in seconds. AI-powered phishing (51%), vishing or voice deepfakes (43%), and general deepfake attacks (41%) are the most concerning new cyberattack techniques for leadership teams, proof of how quickly AI is expanding the attack surface.
Phishing remains the most common and fastest-growing of these threats. Concern has more than doubled over the past year for cybersecurity leaders with director titles and above, rising from 22% to 51%. With generative AI tools, attackers can craft thousands of personalized messages that mimic legitimate communication, turning social engineering into a fully automated operation.
But phishing is only one part of the evolution. Voice cloning and deepfake technology are advancing just as fast, allowing attackers to impersonate trusted figures with unsettling precision. As these tactics spread, human judgment—the last line of defense—becomes increasingly unreliable.
AI Deception Is Going Mainstream
Earlier in 2025, BrightHire CEO Ben Sesser and Pindrop CEO Vijay Balasubramaniyan described fake job candidates who used AI to generate résumés, clone voices, and appear on camera through deepfaked video during remote interviews. Later in the year, an audio deepfake impersonating Secretary of State Marco Rubio targeted foreign ministers, a U.S. governor, and a member of Congress with AI-generated voicemails mimicking his voice.
Even though no recipients were confirmed to have been fooled, the incident illustrates how accessible and convincing these tools have become.
Together, these cases show that AI-driven deception is no longer confined to the inbox. It is entering everyday operations and public discourse, exploiting human trust and procedural blind spots as effectively as it manipulates data. As these manipulative attacks accelerate, AI is transforming the threat landscape faster than most teams can adapt.
The real test now lies in how leaders respond, and hesitation may be the very weakness that leaves their organizations exposed.
The Confidence Crisis: When Silence Becomes a Security Risk
Many cybersecurity executives recognize the risks but falter when it comes to action. 68% say they’re only moderately or somewhat confident in their organizations’ ability to detect and defend against AI-driven attacks in real time. More than half (53%) admit AI is creating new attack points faster than their teams can secure them.
Over one-third (36%) acknowledge that the technology behind modern cyberattacks is more advanced than the tools they have to defend against them. The result is a growing capability gap—attackers are moving faster, automating their tactics, and exploiting defenders’ slower response cycles.
And when confidence erodes, so does transparency. Nearly half (48%) of cybersecurity leaders did not report a material breach to executive leadership or the board in the past year, and 71% say they would consider not reporting an incident at all. It’s a troubling pattern that shows how hesitation can be as dangerous as any technical weakness. But what drives that silence?
Many organizations choose not to disclose incidents that could strengthen their defenses, often out of fear rather than negligence. 44% of cybersecurity leaders fear financial or reputational fallout, and 40% worry about punitive responses from leadership. In this kind of environment, transparency feels risky and silence becomes the safer choice.
Yet every unreported breach hides valuable lessons, leaving the same vulnerabilities exposed for attackers to exploit again.
Underreporting and limited visibility create a false sense of control just as AI-driven attacks accelerate, widening the gap between confidence and capability. Closing that gap will define what resilience looks like in 2026 and beyond.
Rebuilding Confidence in the AI Era
The past year has made one thing clear: AI is not only changing how attacks happen, it is reshaping how organizations think, react, and lead under pressure.
The path forward is cultural as much as technical. To close the confidence gap, leaders should focus on 4 key priorities:
- Train with Intent—Build AI-aware reflexes before crisis hits. 51% of organizations increased security awareness training last year, and 43% have provided training on both generative and agentic AI cybersecurity risks specifically. These programs must move from reactive to routine, embedding muscle memory that kicks in before—not after—a breach.
- Leverage AI Defensively—Use AI to predict, not just respond. 96% of organizations use AI to automate routine tasks, freeing cyber teams for higher-value defense like advanced threat hunting (44%) and upskilling in advanced cyber domains (43%). AI’s real impact lies in empowering people, not replacing them. It can enhance detection speed, visibility, and resilience—if used intentionally.
- Encourage Transparency—Replace blame with learning loops. Visibility and accountability must become non-negotiable. Yet 37% of cyber leaders cite a lack of clear internal reporting protocols or secure channels for disclosing incidents without fear of blame as a key reason underreporting persists. Open, blame-free communication is essential—leaders can’t fix what they never see.
- Collaborate for Scale—Extend your security perimeter through partnerships. 66% of organizations now rely on managed security service providers for extended coverage and advanced AI capabilities. Partnership is no longer a fallback—it’s a force multiplier.
The Road Ahead
Technology alone won’t restore confidence. Leadership will.
As we move into 2026, the focus must shift from reaction to readiness. The enterprises that succeed will be those that combine automation with accountability and speed with transparency.
AI can strengthen defenses, but only if leaders build the trust, training, and clarity to use it effectively.
In an era defined by accelerating threats, confidence and resilience will belong to the organizations that turn hesitation into clarity and silence into action.
