Hello, you are using an old browser that's unsafe and no longer supported. Please consider updating your browser to a newer version, or downloading a modern browser.
AI CERTs built AI+ Ethical Hacker for security people who use AI in testing and defense, and who now have AI systems in scope. This guide, current as of October 2026, covers the 50-question exam, the twelve blueprint domains and their weights, what it costs, how yearly recertification works, the roles it fits, and how it compares with other ethical hacking and AI security certifications.
AI+ Ethical Hacker is AI CERTs' certification for security professionals who bring AI into ethical hacking and defense work. AI CERTs lists it under code AT-220 and now sells it as AI+ Ethical Hacker Practitioner.
It checks that you can run the core of an engagement (reconnaissance, enumeration, vulnerability assessment and penetration testing) with AI helping at each step, and that you can test AI systems themselves, from LLM applications and RAG pipelines to AI agents. The exam is 50 multiple-choice questions in 90 minutes, with a 70% passing score.
The blueprint goes well past offense. It also covers threat hunting, incident response, identity and Zero Trust, and AI governance under the NIST AI RMF. There are no mandatory prerequisites, though AI CERTs recommends networking, operating system and basic programming knowledge. The credential is issued by AI CERTs, which publishes the exam blueprint and sets the one-year recertification rules.
Explain attack surfaces, legal boundaries, testing lifecycles and the frameworks testers use.
Run reconnaissance, enumeration and vulnerability assessment, then rank risk and plan remediation.
Use AI for recon, pen testing, LLM and RAG assessment and reporting, and check what it produces.
Work with SOC operations, threat hunting, forensics, Zero Trust and non-human identity controls.
Spot prompt injection, RAG and agent weaknesses, then disclose and report findings responsibly.
Four design choices set it apart from the ethical hacking and pen testing certifications most testers already know.
AI CERTs pitches this one as different from conventional ethical hacking certifications because it centers on applying AI techniques. In practice the blueprint puts AI on both sides of the engagement. You use AI to speed up reconnaissance, vulnerability analysis and reporting, with a human validating each result. You also test AI itself, including LLM applications, RAG pipelines and AI agents, against the OWASP LLM Top 10 and adversarial machine learning attacks.
Four domains worth 33% of the exam sit on the defense side: threat hunting and SOC operations, incident response and forensics, identity and Zero Trust, and AI governance. The blueprint expects a tester to know how findings land with the blue team.
AI CERTs' scheme sets no required education, experience or prior certification. It recommends operating system, networking and introductory ethical hacking knowledge, plus basic programming with Python preferred.
AI+ Ethical Hacker is valid for one year, a much shorter cycle than the three years PenTest+, SecAI+ and COASP run on. The only way to renew is re-examination: a proctored multiple-choice recertification exam that AI CERTs' Candidate Handbook prices at $95, taken before your certificate expires, as the AI CERTs recertification page sets out.
Know this trade-off before you start. There are no continuing education credits to log, but you sit an exam every year to stay certified. Miss the date without asking for a grace period and the certification is suspended, and you may have to go through the full process again as a new candidate.
The certification, the AT-220 exam, and how renewal works as of October 2026, taken from AI CERTs' program pages, certification scheme, Candidate Handbook and store.
AI CERTs' certification body lists six AI security certifications, and AI+ Ethical Hacker is the one built for ethical hacking practitioners. The store now sells each one under a new name, while the certification body pages still use the older names.
Now sold as AI+ Security Practitioner, Expert and Strategist (AT-2101 to AT-2103). A three-step track that moves from foundational AI-driven security concepts to controls and threat detection, then to strategy, governance and risk management.
Now AI+ Network Practitioner (AT-510). Covers AI applied to networking, including SDN and NFV, AI-powered network automation and orchestration, and network security.
Now AI+ Security Compliance Practitioner (AT-230). Pairs cybersecurity compliance principles with AI, with a focus on compliance processes, risk management and alignment with regulatory standards.
AI+ Ethical Hacker works best in the middle of an offensive security path. The exam has no prerequisites, but AI CERTs recommends networking, operating system and introductory ethical hacking knowledge first, and the blueprint moves fast from those basics into AI-assisted testing. After it, the next step depends on whether you go deeper into offense, AI security or defense.
Networks and security basics
Learn to test, with and against AI
Two questions to settle before you book: can you take the exam, and should AI+ Ethical Hacker be the credential you go for. The first one has an easy answer.
No experience check needed.
AI CERTs sets no mandatory prerequisites, so anyone can test. As of October 2026 its store sells the self-paced program for $495, with about 40 hours of video, an e-book, podcasts, quizzes, labs and the online proctored exam with one free retake. If you prepare another way, the Candidate Handbook lists a $195 exam-only fee. Vouchers, retakes included, expire six months after purchase, so plan your study inside that window.
Same exam, with a live instructor.
AI CERTs lists a five-day instructor-led format, live virtual or in a classroom, delivered through Authorized Training Partners. Training Camp's five-day AI+ Ethical Hacker boot camp is one of those options, taught on AI CERTs' authorized courseware with the AT-220 exam voucher included, so there is nothing extra to buy before you test.
AI+ Ethical Hacker is a newer credential, so it does the most for you on top of hands-on experience or an established security certification. These roles come from the job list AI CERTs publishes for the program:
Runs authorized attacks against applications, networks and cloud systems to find what an attacker would. Domains 3 through 5 follow that work from reconnaissance and enumeration to validated findings and the final report.
Scans, validates and ranks findings so remediation goes where the real risk is. Domain 4 covers CVE, CVSS and CWE, false positives, and using AI to summarize CVEs and draft risk rankings that a person then checks.
Tests web apps and APIs for flaws like SQL injection, XSS, broken authorization and BOLA. AI CERTs names this role directly, and Domain 5 covers web, API and password testing in detail.
Watches alerts, hunts threats and tunes detections in a SOC. Domain 8 covers SIEM, SOAR, EDR and UEBA, hunting methods, and Sigma and YARA rules, with AI used for triage and log summaries.
Runs adversarial exercises that test whether defenders notice and respond. The capstone domain ties the offensive work to detection engineering, incident response and reporting in a single engagement.
Assesses LLM applications, RAG pipelines and AI agents for prompt injection, retrieval poisoning, excessive permissions and data leakage. Domain 6 and the adversarial machine learning material in Domain 7 cover this work.
The certifications people weigh against AI+ Ethical Hacker are a general pen testing credential, an offensive AI security credential, and an AI security credential for defenders. Here is how they line up.
| AI+ Ethical Hacker | CompTIA PenTest+ | EC-Council COASP | CompTIA SecAI+ | |
|---|---|---|---|---|
| Issuer | AI CERTs | CompTIA | EC-Council | CompTIA |
| Exam Format | 50 multiple-choice questions in 90 minutes | Up to 90 questions, multiple choice and performance-based, 165 minutes | 65 multiple-choice + 5 practical, 6 hours | Up to 60 questions, multiple choice and performance-based, 60 minutes |
| Passing Score | 70% (35 of 50) | 750 (scale of 100 to 900) | 70% to 80% (varies by exam form) | 600 (scale of 100 to 900) |
| Prerequisites | None mandatory | None (3 to 4 yrs recommended) | 2 yrs infosec or official training | None (Security+ and CySA+ recommended) |
| Focus | AI-assisted testing and defense, plus LLM, RAG and agent security | Hands-on pen testing, full lifecycle | Attacking and hardening AI systems | Securing AI systems, AI in security operations |
| Exam Cost | $195 exam only, $495 with self-paced course | ~$439 voucher | $650 voucher | $298 voucher |
| Renewal | $95 recertification exam every year | 60 CEUs over 3 years | 120 ECE credits over 3 years | CE over 3 years, or retake |
| DoD 8140 Approved | No claim made | Yes (10 DCWF roles) | No listing claimed | Not listed in Matrix V2.1 |
| Best For | Security practitioners adding AI to testing and defense | Hands-on testers needing an 8140 credential | Pen testers and red teamers adding AI targets | Practitioners defending and governing AI systems |
Prices are as of October 2026. AI+ Ethical Hacker is the only one of the four with an all multiple-choice exam, and you renew it every year instead of every three.
Certification Matchups: the published facts, strengths and weak spots of 33 certifications, side by side.
Our AI+ Ethical Hacker boot camp covers the certification domains in five days (about 40 hours) of live, instructor-led training, in person or live online, as an AI CERTs Authorized Training Partner. Enrollment includes AI CERTs' authorized courseware with videos, e-books and quizzes, hands-on labs in a provided environment, a capstone project, a mobile study app, exam-style practice questions, your AT-220 exam voucher, a free retake if you need one, and 180 days of one-on-one coaching after class.
How AI changes hacking, the AI attack surface, recon skills, and the certifications testers weigh next.
How attackers and defenders both use AI now, what the AI+ Ethical Hacker program covers from AI-driven reconnaissance to protecting AI models, and how it differs from CEH.
The ten ASI risks for AI systems that plan, remember, call tools and act with delegated authority, and how the list differs from the OWASP Top 10 for LLM Applications.
The five PT0-003 domains and their weights, what the performance-based questions look like, what changed from PT0-002, and a study plan that follows the blueprint.
What SecAI+ costs, whether employers are asking for it yet, who it is built for and who should wait, and how it lines up against other AI security certifications.
SYN, connect, UDP, ACK, idle and protocol scans, when each one fits, and the mistakes that produce misleading results. Good groundwork for the reconnaissance and enumeration domains.
What AI does well in security, what it still can't do, and which jobs are emerging because of it, with practical steps for professionals getting ready for the change.
How to read an official exam outline, use domain weights to plan study time, and turn the objectives into a study plan. Useful for a blueprint split twelve ways.
AI CERTs' current blueprint splits the exam into twelve domains, and none of them is larger than 9 percent. Domain 1 is the smallest at 5 percent, seven domains carry 9 percent each, and the last four carry 8 percent each. With weight spread that evenly across 50 questions, there is no domain you can afford to skip. Click any domain for what it covers.
The smallest domain. Confidentiality, integrity and availability, common threats, attack surfaces, hacker roles, Rules of Engagement and responsible disclosure, and the difference between a vulnerability assessment and a penetration test. It introduces the Cyber Kill Chain, MITRE ATT&CK, MITRE ATLAS, the OWASP Top 10 and the OWASP LLM Top 10, then covers checking AI output, hallucination risk and human-in-the-loop validation.
IP and MAC addressing, subnets, ports, TCP and UDP, DNS, HTTP and HTTPS. Linux file systems, permissions, processes, services and logs. Web requests and responses, cookies, sessions, APIs and JSON. It closes with passive and active reconnaissance, OSINT, footprinting, subdomain discovery and AI-assisted reconnaissance, including the risk of leaning on AI too heavily.
A structured reconnaissance workflow that starts from the authorized scope. Port, service, DNS, web, API and banner enumeration, plus cloud storage, metadata, IAM visibility, Kubernetes and container registries. AI is used to cluster assets, build inventories, flag high-risk services, map possible attack paths and draft the reconnaissance report.
CVE, CVSS and CWE, exploitability, false positives and risk ratings. Authenticated and unauthenticated scanning across networks, web applications, dependencies, containers and cloud configurations, followed by manual validation and evidence collection. AI helps summarize CVEs, rank risk and draft reports, with a human checking the result before remediation plans go out.
The penetration testing lifecycle, then the flaws it targets: broken authentication and authorization, SQL injection, XSS, file uploads, sessions, BOLA, mass assignment, rate limiting, password storage, brute force and MFA. It extends to cloud IAM, public storage, container images, Kubernetes and CI/CD secrets, with AI used for payload explanation, test-case generation and report writing.
How LLM applications are built and where they break. The OWASP LLM Top 10, direct and indirect prompt injection, jailbreak attempts and prompt-based data exfiltration. RAG risks such as retrieval poisoning, insecure ingestion and vector database exposure. AI agent tool calling, excessive permissions and shadow agents, plus controls like output validation, DLP, prompt isolation and human approval.
Datasets, features, labels, training and testing. Supervised learning for malware, phishing and login anomaly detection, unsupervised learning for clustering and behavioral baselines, and NLP for threat intelligence and IOC extraction. The adversarial side covers evasion, data poisoning, model extraction and model inversion, with explainable AI methods such as SHAP and LIME.
SOC tooling including SIEM, SOAR, EDR, XDR, IDS, IPS and UEBA. Hypothesis-driven, IOC-based, behavior-based and MITRE ATT&CK-based hunting. Detection engineering with Sigma and YARA rules, tuning and false-positive reduction, and AI for alert triage, log summarization, event correlation and drafting detection rules.
The incident response lifecycle from preparation through lessons learned. Classifying malware, phishing, breach, insider, cloud and AI application incidents. Chain of custody, hashing and disk, memory, log, browser and network forensics. It also covers AI-specific incidents such as prompt injection, model leakage, data poisoning, RAG exposure and agent misuse.
IAM fundamentals, roles and privileges, Zero Trust, least privilege and continuous verification. Privileged Access Management and just-in-time access. Service accounts, API keys, tokens and workloads as non-human identities, common cloud IAM misconfigurations, and identity controls for AI agents.
AI policy, inventories, risk registers, data and model governance, and accountability. The NIST AI RMF functions (Govern, Map, Measure, Manage) and generative AI risk documentation. Privacy, data minimization, bias and transparency, then how to package a finding for responsible disclosure with evidence, severity, remediation guidance and an executive summary.
One engagement in eight phases: scoping and Rules of Engagement, reconnaissance and OSINT, vulnerability assessment, AI-assisted penetration testing, LLM and AI application testing, detection engineering and threat hunting, incident response, and governance and reporting. On the exam, AI CERTs delivers case-study items as scenario-based multiple-choice questions, with no separate project to submit.
Domains and weights reflect the exam blueprint on AI CERTs' AI+ Ethical Hacker Practitioner program and store pages, checked on October 21, 2026. AI CERTs' certification scheme document (version 3.0) still lists an older set of twelve domains under exam code AIC-ETH-101, so confirm the current blueprint with AI CERTs before you book.
The questions candidates ask most often when researching the AI CERTs AI+ Ethical Hacker certification.
AI+ Ethical Hacker is an AI CERTs certification that tests how you apply AI to ethical hacking, penetration testing and cyber defense, and how you test AI systems such as LLM applications and AI agents. Its certification code is AT-220. AI CERTs now sells it as AI+ Ethical Hacker Practitioner, and its store notes that the program was formerly known as AI+ Ethical Hacker.
No mandatory ones. AI CERTs' certification scheme sets no required education, experience or prior certification, and any candidate may attempt the exam. It recommends a basic understanding of computer systems and operating systems, networking fundamentals, introductory ethical hacking concepts, and basic programming awareness, with Python preferred but not required.
The exam has 50 multiple-choice questions and a 90-minute time limit, and you need 70% (35 of 50 correct) to pass. It is taken online with remote proctoring through Proctor 365, AI CERTs' own proctoring platform. The 90 minutes include 5 minutes to read and sign the Candidate Agreement and 5 minutes for the Proctor 365 tutorial. Some questions are scenario-based case studies, still in multiple-choice form.
AI CERTs' current blueprint starts with Cybersecurity, Ethical Hacking, and AI Foundations at 5%. Seven domains carry 9% each: networking and reconnaissance fundamentals, AI-enhanced reconnaissance and enumeration, vulnerability assessment, AI-assisted penetration testing, LLM and RAG security, machine learning and adversarial AI, and threat hunting and SOC operations. The last four carry 8% each: incident response and forensics, identity and Zero Trust security, AI governance and responsible disclosure, and an end-to-end capstone engagement.
As of October 2026, AI CERTs' Candidate Handbook (version 9.0, August 2026) lists the AI+ Ethical Hacker exam-only fee at $195 USD. The AI CERTs store sells the self-paced AI+ Ethical Hacker Practitioner program for $495, which includes the course content and the online proctored exam with one free retake. Training providers, Training Camp included, also bundle the exam voucher with instructor-led training.
AI CERTs certifications are valid for one year from the date of issue. The only way to recertify is to pass a proctored recertification exam before the certificate expires, and the Candidate Handbook lists that exam at $95. If you can't finish in time, you can ask for a grace period before expiry, which the scheme sets at 30 to 90 days. Without one, a lapsed certification is suspended and you may have to start over as a new candidate.
AI CERTs' certification scheme for AI+ Ethical Hacker dates to version 1.0 on January 8, 2024, with revisions on November 28, 2025 and March 18, 2026. The program is now sold as AI+ Ethical Hacker Practitioner under code AT-220, with a twelve-domain blueprint on AI CERTs' program and store pages. The scheme document still lists an older set of domains under exam code AIC-ETH-101, so confirm the blueprint with AI CERTs before you book.
PenTest+ (PT0-003) is a general penetration testing certification with up to 90 questions, including performance-based items, in 165 minutes and a passing score of 750 on a 100 to 900 scale. AI+ Ethical Hacker has 50 multiple-choice questions in 90 minutes and centers on using AI in testing and defense, plus testing LLM, RAG and agent systems. PenTest+ is approved for DoD 8140 and renews every three years with 60 CEUs. AI+ Ethical Hacker makes no 8140 claim and renews every year by exam.
AI CERTs says it operates as an independent certification body under ISO/IEC 17024:2012, and the AI+ Ethical Hacker certification scheme describes its design as aligned with the principles of ISO/IEC 17024. Its accreditation page, last updated in April 2024, says AI CERTs is actively pursuing ANAB accreditation. None of its pages for this certification claim ANAB accreditation or DoD 8140 approval, so if you need a credential for a DoD 8140 work role, check the current qualification matrix on the DoD Cyber Exchange.
AI CERTs lists ethical hacker, penetration tester, vulnerability assessment analyst, cybersecurity analyst, web application security tester, red team associate, and AI security testing associate as job roles for the program. It is a newer credential, so it carries the most weight alongside hands-on experience or an established security certification such as PenTest+ or CEH.
Coding isn't required to sit the exam. AI CERTs' scheme recommends basic programming awareness, with Python preferred but not required, and its program page suggests basic Python, Java or C++ for security automation. You also need working knowledge of networking, Linux, HTTP and web servers, since Domain 2 covers those directly.
AI CERTs offers the official course two ways: five days of instructor-led training through Authorized Training Partners, or about 40 hours of self-paced video, e-book, podcast and lab content. Under the Candidate Handbook, exam vouchers, retakes included, expire six months after purchase, so plan your study inside that window. There is no published recommended study time beyond the course length.
It depends on what you need it for. It fits if you already work in security and want structured coverage of AI-assisted testing, LLM and agent security, and AI in the SOC, in a 90-minute exam with no experience requirement. It fits less well if you need DoD 8140 approval or a performance-based exam that proves hands-on skill, and the yearly recertification exam adds a recurring cost.
If you're weighing AI+ Ethical Hacker for yourself, sorting out funding, or planning AI security training for a team of testers, tell us where you are and we'll help you plan the next step.