Straight answer
Can prompt injection be fixed?
Probably not completely. The NCSC expects prompt injection never to be fully mitigated the way SQL injection can be, so the aim is to reduce likelihood and limit impact.
Last reviewed against NCSC (UK), Anthropic and UK Government (DSIT and Home Office). Sources
The answer
Why a full fix is unlikely
SQL injection can be prevented with parameterised queries. The NCSC says an LLM has no equivalent separation between instructions and data, because it predicts the next token, so prompt injection probably cannot be fully prevented [2]. It suggests treating the tool as an "inherently confusable deputy" [2].
What to do
Reduce and contain
Buying
What to ask suppliers
A red flag
A supplier who says their product stops prompt injection. The NCSC advises caution about any such claim [2]. Ask instead what the tool can access and what controls sit outside the LLM.
Residual risk
Who should accept it
The NCSC asks that security teams and risk owners accept prompt injection as a residual risk that products alone cannot remove [2]. In practice that means a named person signs off the remaining exposure for each AI tool, knowing what the tool can reach and what a hostile message might make it do. OWASP places prompt injection at the head of its 2025 list for LLM applications, so settle it before the lower-ranked items [9].
The full method is on prompt injection defence.
Keep reading
Read next
Am I exposed?
A practical checklist to see whether your payment processes, inboxes, systems and AI tools are open to AI powered attacks, for UK organisations.
Read the guide
How to defend
An ordered plan for UK organisations to defend against AI powered attacks: payment checks, phishing, patching, AI tool limits, logging and the board.
Read the guide
Timeline
A dated, sourced timeline of AI related cyber attack cases and UK guidance, from the Arup deepfake fraud in 2024 to the 2026 breaches survey.
Read the guideHow CyPro can help
Need a hand with any of this?
Patching is one job. Keeping a vulnerable system safe until it is patched, and proving it, is several more. CyPro's consultants, testers and SOC analysts can take any of them off your plate.
Vulnerability scanning
Find every internet-facing system you run, including the forgotten ones, and see which are still on a vulnerable build.
See the scanning service
Hardening and segmentation
Our consultants put the controls in place that limit the damage when something is missed: firewall rules, network isolation and tighter admin access.
More on cypro.co.uk
24/7 security monitoring
Our SOC analysts watch your systems day and night for the signs of an attack in progress: unexpected processes, new outbound connections, logins that do not fit.
More on cypro.co.uk
Penetration testing
A focused test of your systems and the controls around them, so you know they hold before an attacker finds out for you.
More on cypro.co.uk
Incident response
If you may already be compromised, our responders contain it, preserve the evidence, find out what was touched and get you running again safely.
More on cypro.co.uk
Fractional CISO for the board
A senior security lead who advises the board on what to fix first, what to accept and what to tell customers, and owns the decision record.
More on cypro.co.uk
Know your exposure
Not sure you are protected?
Many of the risks here sit in how an organisation uses AI: what tools staff run, what they can reach and who owns the risk. AI governance work sets that out so controls match the tools in use.