Work and careers · The ruling
Can you use AI to do cyber security work instead of people?
Verdict
The verdict
Not on current evidence. UK sources show AI changing cyber work and raising demand for AI skills, with the sharpest squeeze on entry-level hiring. None of them forecasts job numbers.
No means: no, or not in a way you should rely on. How we rule

Short answer: the UK sources describe AI being used on both sides of cyber security, with demand for AI skills going up. The pressure they record falls hardest on people trying to get their first job. Anyone quoting you a firm percentage of cyber jobs that AI will remove is guessing, because no primary source we read gives one.
The top search results for this question are a Reddit thread, a Quora answer, vendor blogs and a LinkedIn post. They are all opinion and none of them cites anything. So we read what the National Cyber Security Centre (NCSC), the Department for Science, Innovation and Technology (DSIT) and the Information Commissioner’s Office (ICO) have actually published.
What the UK sources say about AI and cyber security work
The main labour market evidence is DSIT’s annual report, Cyber security skills in the UK labour market. The 2026 edition was published on 29 September 2026 and mostly covers 2025. It counts approximately 145,900 people in the UK cyber security workforce, up about 2% on the previous study, and notes that growth is slowing. Core cyber security job postings rose 7% in 2025, after the 2025 edition had recorded a 33% fall in 2024.
In the 2026 edition, 70% of cyber security businesses said staff were using AI in their day-to-day work, up from 53% the year before. AI skills were explicitly asked for in 9% of core cyber postings, up from 4%, and automation skills appeared in 15% of postings, up from 11%. Interviewees widely expected that AI and automation will transform the sector. That describes a field changing what it hires for. It does not describe one shrinking because of AI.
The skills shortage has not gone away either: over half (57%) of businesses had a basic technical skills gap, up from 49% the year before.
Which cyber tasks AI already does well
The NCSC’s 2024 assessment says plainly that AI can improve the detection and triage of cyber attacks. Triage means sorting a large stream of alerts so a person sees the ones that matter, and that kind of high-volume, repetitive pattern-spotting suits software. The NCSC’s 2025 assessment adds that AI will also aid system owners and software developers in securing systems.
Be careful how far you take this. Neither source gives a list of tasks that AI now does from start to finish, and neither measures how much work it has taken off people. Vendor claims about specific tools are a different kind of evidence, and we have left them out here.
Doing a task well is also different from owning the result. A tool can flag a thousand alerts. It has not decided which three to act on, and it cannot explain that choice to a board or a regulator.
How attackers using AI changes demand for defenders
The NCSC’s January 2024 assessment judged that AI will almost certainly increase the volume and heighten the impact of cyber attacks over the following two years. It found the main gains for attackers were in reconnaissance and social engineering. It also said this impact will be offset by the use of AI to enhance cyber security resilience, though more work is needed to measure how much.
Its May 2025 assessment, looking to 2027, goes further. It warns that the time between disclosure and exploitation has shrunk to days and AI will almost certainly reduce it further. It expects a digital divide between systems that keep pace with AI-enabled threats and a large proportion that are more vulnerable.
The NCSC does not turn this into a jobs forecast. Our reading is that faster, cheaper attacks put more pressure on defence. Whether that leads to more jobs, fewer jobs or different ones is not something the assessments say.
Entry-level roles: where the real pressure is
This is where the data is most worrying. DSIT found that demand for entry-level applicants with under one year of experience remained low in 2025, at 16% of postings. The 2025 edition put the figure at 25% in 2022. Employers interviewed for the 2026 report were concerned that the pathway to becoming an experienced cyber security professional was narrowing.
The report does not blame AI for that fall, so neither should you. It is still a fair question, though. If routine triage is increasingly done by tools, juniors need other ways to build the judgement that senior people rely on.
How to do it well
If you are starting out, it helps to:
- Learn why an alert matters as well as how to clear it, and practise explaining your reasoning in plain English.
- Treat AI skills as part of the job. Employers are asking for them more often.
- Look at the UK Cyber Security Council’s professional titles and career pathways and see which route fits you.
- Write your applications honestly. DSIT heard from employers that AI-generated CVs make candidates harder to assess. Our pages on using AI to write your CV and applying for jobs cover where AI help is fine and where it backfires.
- Be careful what you paste into a chatbot. Work logs and incident details may hold personal or confidential data.
When you read a forecast about AI and jobs, check it yourself:
- Check who is saying it. A vendor selling an AI security product has a reason to talk up automation, and a training provider has a reason to talk up demand. That does not make them wrong, but it is worth knowing.
- Find the original document. Read the exact sentence. Does it say what the headline says, or something close to it?
- Check the year. DSIT’s 2026 report mostly describes 2025. A figure about last year is a description, not a forecast.
- Look for the method. A claim with no sample, no method and no date is an opinion.
- Compare editions. Year-on-year changes, like entry-level demand falling from 25% in 2022 to 16% in 2025, tell you more than any single number.
If you use an AI assistant to summarise a report for you, check its summary against the source.
Rules and risks
Accountability is the clearest task that still needs a person. Under UK GDPR the security principle (Article 5(1)(f)) requires personal data to be processed securely, and Article 32 requires appropriate technical and organisational measures. The ICO’s guidance says that as a data controller you are responsible for compliance, including what your processor does with the data. It also expects someone to have day-to-day responsibility for information security. If an AI tool misses a breach, the organisation still answers for it.
The profession is also moving towards named, accountable individuals. The UK Cyber Security Council was established by Royal Charter, and the NCSC describes it as setting and maintaining the standard for professional competence in the profession. Chartered Cyber Security Professional is its highest level of recognition. Professional titles attach to people. They do not attach to software.
The NCSC’s guidelines for secure AI system development point the same way. They ask AI providers to take ownership of security outcomes for customers and to embrace radical transparency and accountability. If your organisation brings in an AI security tool, someone still has to assess it, configure it and watch it.
Live incidents are the other case. Deciding whether to pull a system offline, what to report and to whom involves incomplete information, business cost and legal duties. That is a judgement call. None of the sources we read suggests handing it to a model.
What the sources do not settle
A lot. None of them forecasts how many cyber security jobs there will be in five years, or how many AI will remove. DSIT’s 2026 report dropped its workforce gap estimate altogether because of data limitations and lower demand, so even the size of the shortage is now less certain. The NCSC assessments are about threat, not employment. We found no UK primary source that gives a job-loss percentage for this field.
The honest answer to “will AI take my cyber job?” is that nobody can say with confidence. A confident wrong answer is a worse thing to plan a career around than an honest “we don’t know yet”.
Sources
7- The near-term impact of AI on the cyber threat (January 2024), NCSC
- Impact of AI on cyber threat from now to 2027 (May 2025), NCSC
- Cyber security skills in the UK labour market 2026, DSIT
- Cyber security skills in the UK labour market 2025, DSIT
- A guide to data security (UK GDPR), ICO
- About the UK Cyber Security Council, NCSC
- Guidelines for secure AI system development, NCSC
FAQ
Questions people ask
Can AI replace cyber security analysts?
No UK source we read says so. The NCSC says AI can improve the detection and triage of cyber attacks, which is part of an analyst's job. UK GDPR still makes the organisation responsible for security, and someone has to own the decisions. No source forecasts how many analyst roles there will be.
Is cyber security still a good career with AI?
The DSIT 2026 report counts about 145,900 people in the UK cyber security workforce, up about 2% on the year before. Core job postings rose 7% in 2025, and 57% of businesses had a basic technical skills gap. Growth is slowing, though, and entry-level demand is low, so check the latest edition before you decide.
Which cyber security tasks can AI automate today?
The NCSC names detection and triage of cyber attacks as areas where AI can help. DSIT found 70% of cyber security businesses reporting that staff use AI in their day-to-day work. Neither source lists exactly which tasks are fully automated, so be wary of anyone who claims a precise list.
Does AI make cyber attacks more likely, and does that create more jobs?
The NCSC judged in 2024 that AI will almost certainly increase the volume and heighten the impact of cyber attacks. In 2025 it warned of a digital divide between systems that keep pace with AI-enabled threats and those that do not. It does not say whether this means more defender jobs.
Who is accountable when an AI tool makes a security mistake?
The organisation. ICO guidance says a data controller is responsible for complying with UK GDPR, including what its processors do with the data, and Article 32 requires appropriate technical and organisational security measures. Using an AI tool does not move that duty onto the tool.
General information for UK readers, not legal, financial or medical advice. How verdicts are decided.