Retaining defensive advantage in the age of frontier AI cyber capabilities
As AI accelerates vulnerability discovery, organisations must raise their security baselines to safeguard their cyber security.
Frontier AI: what you need to know
The enhanced threat from AI-powered cyber attacks is real, but organisations can use AI to improve defences and embrace opportunities.
Artificial intelligence (AI) is no longer a distant or speculative issue for cyber security. The most advanced AI tools, often referred to as frontier AI, make it easier, faster and cheaper for even low-skilled hackers to carry out sophisticated cyber attacks.
Key messages for leaders and board members
-
AI is making it easier and faster for criminals to attack those organisations that do not have basic cyber security protections in place.
-
The NCSC believes that AI will ultimately be a ‘net positive’ for cyber security, but the path to reaching this point requires urgent action now and board-level engagement.
-
The rise of agentic AI tools (a new class of frontier AI) means organisations need to understand how they are used, and what access they have to systems and data.
-
Maintaining good cyber security fundamentals remains the best way to protect your organisation from all digital threats, whether those are AI-assisted or not.
What does frontier AI mean for cyber security?
Cyber attacks which once required specialist skills (such as writing code, understanding system architecture, or discovering weaknesses in computer systems) can increasingly be automated using AI. This does not fundamentally change cyber risk, but it does increase the speed and scale at which existing weaknesses can be found and exploited.
Agentic AI tools - a new class of frontier AI - go even further and can plan, make decisions and take actions on your behalf.
However, by carefully adopting frontier AI tools, network defenders can retain an advantage over cyber criminals. Technology suppliers are already using AI to identify and fix vulnerabilities in their products and services to protect users from new threats.
Organisations that act now can strengthen their resilience, reduce future disruption and demonstrate to customers, prospects and suppliers that they’re ready to face the future opportunities that AI promises. Those that delay risk exposing their organisation to a wave of vulnerabilities that AI tools are rapidly exploiting.
Want to know more about AI?
The NCSC provides guidance to help organisations keep pace with AI developments.
Thinking carefully before adopting agentic AI
When it comes to using agentic AI, make sure you can walk before you run.
10 questions to ask when using AI models to find vulnerabilities
Using Artificial Intelligence to find vulnerabilities can bring added security considerations.
Impact of AI on cyber threat from now to 2027
An NCSC assessment highlighting the impacts on cyber threat from AI developments between now and 2027.
AI: a deeper look
The following resources have been created by the NCSC for security professionals and the wider technical community.
One does not simply defend agentically
Defenders can’t use AI in the same way attackers can, but there’s much they can do to unlock the potential of agentic cyber defence.
The hidden risks of shadow AI
Understanding why staff use unapproved AI tools is key to managing the security challenges they can create.
Managing the cyber risk of agentic AI
Use safeguards, sandboxing and active oversight to realise the benefits of autonomous systems while limiting the unintended activity.
The 'vibe coding spectrum' approach to AI-assisted software development
Different code deserves different levels of oversight, so calibrate your approach to ‘vibe coding’ accordingly.
Preparing for a ‘vulnerability patch wave’
Organisations must act now to prepare for a wave of patches that will address decades of technical debt.
Why cyber defenders need to be ready for frontier AI
Understanding the threats and staying ahead of the adversary
Vibe check: AI may replace SaaS (but not for a while)
If ‘vibe coding’ disrupts the software market like SaaS did 20 years ago, what does this mean for cyber security?
Understanding adversarial attacks against Machine Learning and AI
Introducing a common language to improve awareness, threat modelling, and collaboration on AI security
Prompt injection is not SQL injection (it may be worse)
There are crucial differences between prompt and SQL injection which – if not considered – can undermine mitigations.
International collaboration
Frontier AI is a global challenge that requires a global response. Together with international partners across government, industry and academia, we have contributed to the following publications.
-
Guidelines for secure AI system development
Agencies from 18 countries, including the US, endorse new UK-developed guidelines on AI cyber security.
-
Careful adoption of agentic AI services
Joint guidance, co-authored by the NCSC with international partners, that explores security challenges and risks posed by agentic AI.
-
Securing Artificial Intelligence (SAI); Guide to Cyber Security for AI Models and Systems
Technical report published by the European Telecommunications Standards Institute (ETSI).
Securing Artificial Intelligence (SAI); Guide to Cyber Security for AI Models and Systems
-
Content Credentials: Strengthening Multimedia Integrity in the Generative AI Era
Content Credentials’ guidance, co-authored with the NSA, seeks to counter the erosion of trust.
Content Credentials: Strengthening Multimedia Integrity in the Generative AI Era
-
The AI shift in cyber risk: why leaders must act now
A call to action from the Five Eyes security agencies. The evolving landscape of artificial intelligence (AI) is rapidly transforming cyber risk, and we must act swiftly to remain ahead.