
Will AI Replace Cybersecurity professionals?
AI agents from CrowdStrike, Microsoft, Google and XBOW already triage alerts and find real vulnerabilities. Federal projections still show fast job growth, but entry-level analyst work is where the pressure lands.
No, AI is not on track to replace cybersecurity professionals as a group, and the federal outlook points toward growth. The pressure falls on entry-level work, where AI agents now handle much of the alert triage and phishing review that once trained new analysts.
That is our read of the evidence below. The deployments are real, the job projections are strong, and the gap between those two facts shapes who gets hired next.
Where AI is already doing security work today
The clearest cases come from vendors that sell to security operations centers (SOCs), plus one AI system that competes with human bug hunters.
CrowdStrike: layoffs and Charlotte AI
In May 2025, CrowdStrike said it would lay off 500 employees, about 5% of its workforce, and its CEO tied the move to advances in AI. The company framed AI as something that flattens its hiring curve and speeds product development.
The cuts were not aimed at security analysts. They fell in the front and back office, and CrowdStrike said it would keep hiring "primarily in customer-facing" and product engineering roles.
In June 2025, CrowdStrike unveiled Charlotte AI Agentic Response and Charlotte AI Agentic Workflows, which it described as "marking a leap forward in modern SOC operations." By December 2025, the company reported that Charlotte AI had achieved FedRAMP High Authorization, which clears it for sensitive federal use.
Microsoft: phishing triage agents
In March 2025, Microsoft announced Security Copilot agents, saying its phishing triage agent "can handle routine phishing alerts and cyberattacks." Microsoft's changelog shows the Phishing Triage Agent in Defender entered public preview in July 2025, with a customer spotlight added in September 2025.
Microsoft pitches the tool as "a force multiplier for SOC analysts" that clears user-submitted phishing reports. The agent resolves false positives and escalates only malicious cases that need human expertise.
Google: Big Sleep finds zero-days
In November 2024, Google's Big Sleep agent found a zero-day in the SQLite database engine. Google called it "the first public example of an AI agent finding a previously unknown exploitable memory-safety issue" in widely used software.
In 2025, working from Google Threat Intelligence leads, Big Sleep found CVE-2025-6965. Google said: "We believe this is the first time an AI agent has been used to directly foil efforts to exploit a vulnerability in the wild." Humans did the first half of that job. Google's threat intelligence group was "able to identify artifacts indicating the threat" actors were preparing an attack before the AI pinned down the flaw.
XBOW: an AI on the bug bounty leaderboard
In June 2025, XBOW became the first autonomous system to reach #1 on HackerOne's US leaderboard. It submitted over 1,000 vulnerability reports in a few months, and XBOW head of security Nico Waisman wrote, "All findings were fully automated." Human staff still reviewed reports before submission to meet HackerOne policy. Investors followed: XBOW raised $75 million in Series B funding.
Myths vs reality for security careers
| Myth | Reality |
|---|---|
| AI security tools mean fewer analysts overall. | BLS projects information security analyst jobs to grow from 192,900 in 2025 to 233,400 in 2035. |
| AI agents now run the SOC alone. | Microsoft's triage agent escalates malicious cases to humans. |
| AI-generated vulnerability reports are always useful. | The curl project ended its HackerOne bug bounty after a flood of low-quality AI reports. |
| CrowdStrike's AI layoffs hit its analysts. | The cuts fell in front and back office roles. |
What the numbers say about security jobs
The Bureau of Labor Statistics projects a +21.0% change in information security analyst employment over 2025 to 2035, against 3.5% for all occupations. It expects 14,100 openings a year on average. The median annual wage was $129,180 in 2025, and the typical entry credential is a bachelor's degree.
Surveys from ISC2, the main professional body, add the texture BLS lacks. In July 2025, ISC2 reported that 30% of cybersecurity professionals are already using AI tools. In the same research, 52% said AI will significantly or somewhat reduce the need for entry-level staff.
A year later, the prediction looked closer to fact. In an ISC2 survey published in July 2026, 56% of security professionals said AI had somewhat or significantly reduced the need for entry-level positions. In another ISC2 workforce survey, 44% said their organizations are reconsidering roles and skill needs in response to AI.
We think both signals are true at once: total demand grows, while the first rung of the ladder gets narrower.
Security skills gaining value as AI spreads
- Threat intelligence. Big Sleep's biggest win started with human analysts spotting attacker staging artifacts. Knowing where to point the AI matters.
- Validating AI findings. XBOW's automated reports still went through human review before submission. Triage of machine output is now a job skill.
- Securing AI systems. NIST's draft Cyber AI Profile starts with securing AI systems as one of three focus areas.
- Running AI-augmented teams. The NIST draft's defend area centers on augmenting cybersecurity teams with AI agents.
- Network monitoring with AI tools. ISC2 respondents named network monitoring among the top two areas where AI will raise efficiency fastest, so analysts who tune those tools gain leverage.
The messy middle: hallucinations, slop and draft rules
AI fails in ways that create work for defenders. When attackers misused Claude in an espionage campaign, Anthropic said the model "occasionally hallucinated credentials or claimed to have extracted secret information that was in fact publicly-available." Anthropic called such errors "an obstacle to fully autonomous cyberattacks".
CrowdStrike's own SEC filings acknowledge significant risks associated with AI, even as the company leans on it. An XBOW researcher, Dolan-Gavitt, cautioned that AI pen testers are not a panacea.
The curl project shows the cost of bad AI output. Curl first required reporters to disclose generative AI use. Maintainer Daniel Stenberg later noted the bounty generated seven submissions in one week, and the project shut the program after junk reports that included fabricated vulnerabilities.
Rules lag the tools. NIST released its preliminary draft Cyber AI Profile on December 16, 2025, and as of September 7, 2026 it remained in draft status. KPMG notes the voluntary profile may become a benchmark for regulators judging security diligence. The open question is who trains tomorrow's senior analysts if junior triage work disappears. ISC2 CEO Scott Beale urged security leaders to protect entry-level roles.
Our verdict: AI reshapes security work, with juniors most exposed
We think cybersecurity professionals are among the safer bets in tech. AI already handles phishing triage, false positives and some vulnerability discovery, but every documented deployment keeps humans on escalation, review or intelligence. The risk is concentrated: if your plan relies on years of alert triage to learn the trade, that path is shrinking.
Questions people ask
Will AI replace cybersecurity jobs?
BLS projects information security analyst jobs to grow from 192,900 in 2025 to 233,400 in 2035, a +21.0% change. Our read is that AI changes the work rather than removing the profession.
Is AI hurting entry-level cybersecurity jobs?
In an ISC2 survey published in July 2026, 56% of security professionals said AI had somewhat or significantly reduced the need for entry-level positions. ISC2 CEO Scott Beale urged leaders to protect those roles.
How many cybersecurity professionals use AI tools?
ISC2 reported in July 2025 that 30% of cybersecurity professionals are already using AI tools. In a separate ISC2 survey, 44% said their organizations are reconsidering roles and skill needs because of AI.
Can AI find vulnerabilities on its own?
Yes, in some cases. Google's Big Sleep found a zero-day in SQLite, and XBOW reached #1 on HackerOne's US leaderboard, though humans reviewed XBOW's reports before submission.
What does a cybersecurity analyst earn?
BLS lists a median annual wage of $129,180 for information security analysts in 2025. The typical entry education is a bachelor's degree, with 14,100 openings a year on average.
Sources
Every number, quote and example on this page was checked by script against the sources below before publishing: quotes must match the source word for word, and figures must appear in a source or in BLS data. Our verdict is editorial judgment. Spotted an error? Tell us.
- U.S. Bureau of Labor Statistics: Employment Projections, Information security analysts (15-1212)
- CrowdStrike announces 5% job cuts, says AI is ‘reshaping every industry' (nbcmiami.com)
- CrowdStrike announces 5% job cuts, says AI is 'reshaping every industry' (cnbc.com)
- CrowdStrike cuts 500 jobs in AI pivot, but flags risks (csoonline.com)
- CrowdStrike cuts 500 jobs, cites AI (computing.co.uk)
- Document (sec.gov)
- Document (sec.gov)
- Microsoft unveils Microsoft Security Copilot agents and new protections for AI | Microsoft Security Blog (microsoft.com)
- What’s new in Microsoft Defender XDR at Secure 2025 | Microsoft Community Hub (techcommunity.microsoft.com)
- What's new in Microsoft Security Copilot? (learn.microsoft.com)
- Security Copilot for SOC: bringing agentic AI to every defender | Microsoft Community Hub (techcommunity.microsoft.com)
- Google’s AI Tool Big Sleep Finds Zero-Day Vulnerability in SQLite Database Engine (thehackernews.com)
- Google says ‘Big Sleep’ AI tool found bug hackers planned to use (therecord.media)
- A summer of security: empowering cyber defenders with AI (blog.google)
- AI Bug Hunter Sets Milestone By Claiming Top Spot on HackerOne's Leaderboard (techrepublic.com)
- XBOW's AI reached the top ranks on HackerOne, and now it has $75M to scale up - Help Net Security (helpnetsecurity.com)
- XBOW Raises Series B to Advance Autonomous Penetration Testing | XBOW (xbow.com)
- An AI-Driven Pen Tester Became a Top Bug Hunter on HackerOne (darkreading.com)
- AI Won't Wipe Out Entry-Level Cybersecurity Jobs (darkreading.com)
- ISC2 Research Reveals Cybersecurity Teams Are Taking a Cautious Approach to AI Adoption (isc2.org)
- Hiring for the AI Era: A New Challenge for CISOs (infosecurity-magazine.com)
- ISC2 Survey: 30% of Cyber Pros Using AI Security Tools (isc2.org)
- NIST Issues Preliminary Draft of Cyber AI Profile, a Framework Poised to Alter Security Operations in the AI-Driven Threat Landscape - Lexology (lexology.com)
- NIST Publishes Preliminary Draft of Cybersecurity Framework Profile for Artificial Intelligence for Public Comment (insideprivacy.com)
- Draft NIST Guidelines Rethink Cybersecurity for the AI Era (nist.gov)
- NIST AI Cybersecurity Framework Profile - United States (regulations.ai)
- Cybersecurity: NIST Draft Cybersecurity Framework for AI (kpmg.com)
- Chinese spies used Claude to break into critical orgs (theregister.com)
- Disrupting an AI-orchestrated cyber espionage campaign (anthropic.com)
- Curl ending bug bounty program after flood of AI slop reports (bleepingcomputer.com)
- Curl shutters bug bounty program to stop AI slop (theregister.com)
- cURL Ends Bug Bounty Over AI-Generated Spam Reports - Technology Org (technology.org)
- Curl creator mulls nixing bug bounty awards to stop AI slop (theregister.com)