This article is the research in behind my talk “Entry-Level Cybersecurity in the Age of AI: What Changed and What Comes Next” present at Geek Week Atlanta 2026.
Yup this is another article about AI. Unfortunately, it’s something we interact with more and more every day. So, we have to talk about it. Since I’m new-ish to cybersecurity I started looking into how AI is affecting the Entry-Level positions in the cybersecurity job market. Honestly, it feels like it’s taking over. For example, Anthropic released information about the Claude Mythos Preview and it found bugs that are 20+ years old. As I am someone who wants to be on the ethical hacking side in the future, it makes me rethink my career path. To come back to Entry-level positions, AI is hitting it hard and fast, and we need to find a solution.
Where are we now?
Currently there are 4.8 million unfilled cybersecurity roles globally [1][2]. When it comes to the US, only 19% of those roles are asking for 0 to 3 years of experience [2]. On top of that, we have skill gaps that are growing year after year within the industry [1][3]. So, what does this mean? If we aren’t hiring juniors today, we’re not going to have seniors in the future. Now, let’s talk about it.
How is AI impacting the cybersecurity profession?
As of ISC2’s 2025 report, there are 95% of organizations that report one or more cybersecurity skill gaps [3]. The skill gaps include areas like AI, cloud computing security, risk assessment, and more. These gaps lead to security consequences. ISC2’s report mentions 88% of the organizations have had security incidents from these skill gaps. These numbers are something that as an industry we need to address. The question becomes where do these gaps come from and how do we fix it?
Where do the gaps start?
Education. This is where my perspective comes into play. I have seen education before COVID and after COVID. I completed high school when everything was still handwritten and my grade 12 finals were half my grade. Currently, I’m getting a bachelor’s degree at university. I’ve noticed the education system has changed. There are three things I want to address:
- AI Use on Assignments
- The Cybersecurity Degree
- University Exams
First, AI use on Assignments. Students do use AI on their assignments. In high school, 1 in 2 teens use AI on their schoolwork [4]. In universities, I’ve found research anywhere from about 50% up to about 90% of students use it [5][6]. A lot of this relies on self-reporting so that’s where the data is not exact. Then on top of that, from my own experience, we have professors who may or may not know when AI is used. Some professors will see the em dash and immediately think AI was used. Other professors run it through detection software and go by whatever answer it says. The problem with that is there’s false flags. Especially when writing becomes more technical and jargon heavy. Then AI use on coding assignments… there’s currently no reliable detection that exists.
// So… is this bad?
Well… it depends. If you enter a prompt of “Do this for me,” well that is one hundred present cheating. However, if you use it to assist you to study for exams, assist in research, ask how-to questions, or help check grammar. That is less likely to be seen as bad, as AI was used as a tool to learn.
Next, we have the cybersecurity degree. There are a lot of concepts to cover in cybersecurity. The problem is the degree can’t cover them thoroughly enough. You become a jack of all trades but a master at none. Since this is the case employers value certificates and experience more. Seeing only a degree versus someone with labs, certificates, and a GitHub gives the advantage to the person without the degree as they are willing to learn on their own.
Lastly, examinations. This is something that’s seen across multiple tiers of universities. It doesn’t matter if it’s a Top end tech school like Georgia Tech or a community college. The examinations need improvement. What do I mean by this? A lot of the exams done for cybersecurity are open book/open note. While this can be done well by adjusting time limits or applied questions. Coming from my own experience and research, you currently see questions that copy paste out of the textbook. This is likely due to how the exams are given, which are typically completed online using anti-cheat software. While this software helps, it doesn’t prevent all cheating (Disclaimer: I do not condone cheating in any way and do not do this myself). What do I mean? Well yes, you do scan the room and have the camera on, but what is not seen what is physically on your laptop keyboard/ screen (ex. Sticky notes). Another thing that’s not checked is the location of your phone. So, if the students are using notes or cheating, what are they even learning?
How can we fix this?
Well, there’s three areas we can address:
- Academic Changes
- Adding in new certification pathways
- Training and onboarding strategies
Academically we can change a lot. First, adjust the exams. Remove online and bring back in-person examination. If you use open book exams, improve it to have a shorter time limit and have applied questions. Next embrace AI have students use it. This is something I’ve started to see but get students to use it on their assignments and source it. For example, give the students a research paper then have them use AI to assist them in researching sources. Another example is to have students do an assignment then upload it to AI and see what changes it makes (Note: always remove PII). Then have the students document what they changed or didn’t change from the AI response. Lastly, split the degree. Have two or three different tracts for a student to choose from. There could be a Cybersecurity Management degree for those who want to peruse towards the CISO route. Then have a Cybersecurity Technical degree for those who want toward red or blue teams.
Currently, we have a lot of knowledge based entry-level cybersecurity certificates. Let’s create a new certificate that’s hands on. For example, one known certificate that’s hands on is OSCP. You have 24 hours to hack so many labs then the next 24 hours you write your reports. What if we had this but more entry-level style (of course shorter in length)? Like yes Comptia Security+ kind of hits this with the performance based questions. However, what if the whole exam is like that? What if you had a few mini labs and the test taker has to solve the problem, enter the ticket notes, and mark the ticket as closed. You could also add in a lab that they may or may not be able to solve but can mark as an escalation. Some lab ideas could be remediating a malicious software from a device or analyzing Wireshark packet. Creating a new certificate shows the employer that the person has the ability to solve entry-level problems.
Lastly, adjusting our training/ onboarding strategies. Currently 35% of organizations budget for staff development. If companies all adopt a budget for staff development that can improve those gaps and help prepare employees for more advanced skills as AI develops. Next companies can adopt an apprenticeship model. While some companies may already do this, applying this gives the employee the ability to shadow/ complete tasks with a mentor. This can help the employee learn and grow without making costly mistakes. Another option is adding training bootcamps to the organization. When the employee is hired, have them complete a training course to fast-track entry-level knowledge. While there are currently some training bootcamps most are not well adopted (Third-party, non-accredited, etc.). This may be a costly option, but having a custom training camp can improve some of the knowledge gaps entering the cybersecurity field.
Stat References
[1] ISC2, “ISC2 2024 Cybersecurity Workforce Study,” ISC2, Oct. 2024. [Online]. Available: https://www.isc2.org/Insights/2024/10/ISC2-2024-Cybersecurity-Workforce-Study
[2] CyberSeek, “CyberSeek: Cybersecurity Supply/Demand Heat Map,” National Initiative for Cybersecurity Education (NICE), Burning Glass Technologies, and CompTIA. [Online]. Available: https://www.cyberseek.org/index.html
[4] ISC2, “ISC2 2025 Cybersecurity Workforce Study,” ISC2, Dec. 2025. [Online]. Available: https://www.isc2.org/Insights/2025/12/2025-ISC2-Cybersecurity-Workforce-Study
[5] M. Anderson et al., “How Teens Use and View AI,” Pew Research Center, Feb. 24, 2026. [Online]. Available: https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/
[6] BestColleges, “56% of College Students Used AI on Assignments,” BestColleges Research, 2024. [Online]. Available: https://www.bestcolleges.com/research/most-college-students-have-used-ai-survey/
[7] Digital Education Council / Campus Technology, “Survey: 86% of Students Already Use AI in Their Studies,” Campus Technology, Aug. 28, 2024. [Online]. Available: https://campustechnology.com/articles/2024/08/28/survey-86-of-students-already-use-ai-in-their-studies.aspx
About the Data
Workforce & Industry Stats*
ISC2 2024 Cybersecurity Workforce Study:
https://www.isc2.org/Insights/2024/10/ISC2-2024-Cybersecurity-Workforce-Study
ISC2 2024 Workforce Study — First Look:
ISC2 2025 Cybersecurity Workforce Study:
https://www.isc2.org/Insights/2025/12/2025-ISC2-Cybersecurity-Workforce-Study
AI Use in Education*
Pew Research Center — How Teens Use and View AI (Feb. 2026)
https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/
RAND — Students Using AI for Homework (Mar. 2026)
https://www.rand.org/pubs/research_reports/RRA4742-1.html
Wiley — AI Has Hurt Academic Integrity (2024)
Inside Higher Ed — Students & Professors Expect More Cheating (2024)
Inside Higher Ed — College Students’ Views on AI (2025)
https://www.insidehighered.com/news/students/academics/2025/08/29/survey-college-students-views-ai
Gallup / Lumina — AI Is Routine for College Students (2026)
https://news.gallup.com/poll/704090/routine-college-students-despite-campus-limits.aspx
BestColleges — 56% of College Students Used AI on Assignments
https://www.bestcolleges.com/research/most-college-students-have-used-ai-survey/
BestColleges — 60% of Online Students Report Using AI (2025)
https://www.bestcolleges.com/research/college-student-attitudes-on-ai/
College Board — 69% of High School Students Use AI (2025)
Digital Education Council — 86% of Students Use AI in Studies
AllAboutAI — AI Cheating Statistics
https://www.allaboutai.com/resources/ai-statistics/ai-cheating-in-schools/
University of Reading — 94% of AI Exam Submissions Undetected
https://nerdynav.com/chatgpt-cheating-statistics/
AI Detector Accuracy*
MDPI Information — Evaluating AI Detection Tools in Higher Education (Oct. 2025, peer-reviewed)
https://www.mdpi.com/2078-2489/16/10/905
HackerRank — Traditional Plagiarism Tools Inadequate for AI Code (2025)
University Exam Policies*
Georgia Tech Online MS Cybersecurity — HonorLock Proctoring Policy
https://pe.gatech.edu/degrees/cybersecurity/student-resources
Penn State World Campus — Proctored Exam Policy
https://student.worldcampus.psu.edu/course-work-and-success/proctored-exams
* Stats in these sections reflect a broad reading across multiple sources combined with personal experience. Individual figures are cited inline using IEEE format in the References section above.
The Cozy Hacker 