Dark AI: A Dangerous Accelerant?



Overview

Discover the impact of artificial intelligence on offensive cyber. In this online lecture, we survey how large language models (LLMs) are being used by malicious actors to aid their hacking operations. We consider the availability of such models, guardrail circumvention, and the dual-use nature of synthetic media (deepfakes). Finally, we examine the possibility of fully autonomous offensive cyber capabilities. At the end of this lecture you will understand how rapid AI evolutions are impacting hacking. 

This lecture is part of the 'National Security and Cyber: Hackers, Artificial Intelligence and Surveillance' lecture series, taking place on Tuesdays from 16 September to 21 October 2025. You can choose to register for individual lectures or may choose to register for the entire lecture series at a reduced price.

Due to the nature of the contents of this lecture, it is not suitable for people under 18 years old. 

Please note: this lecture will close to enrolments at 23:59 BST (UTC+1) on 27 September 2025.

Programme details

6.15–7.30pm BST (UTC+1)

This lecture will last approximately 1 hour, followed by questions.

Please join in good time before each lecture to ensure that you have no connection problems. We recommend joining 10-15 minutes before the start time.

Fees

Description Costs
Event fee £15.00

Funding

If you are in receipt of a UK state benefit or are a full-time student in the UK you may be eligible for a reduction of 50% of tuition fees.

Concessionary fees for short courses

Tutor

Dr Craig Jarvis

Tutor

Craig is an independent academic specialising in national security and technology. He lectures at Oxford, Cambridge, Imperial College London, and Royal Holloway. His next book, ‘Cyber Terrorism: Extremism & Hacking’ will be published in 2026. Previous works include 'CryptoWars: The Fight for Privacy in the Digital Age' and contributions to 'Next Generation Enterprise Security'. His research also appears in peer-reviewed journals, including 'Intelligence and National Security'.
Craig holds a PhD in cybersecurity and history and has earned master’s degrees in international security, information security, and classical music, studying at Oxford, King’s College London, and Royal Holloway.
With extensive industry experience, his last permanent role was Chief Technology Officer at DXC Security, then the world’s largest security services provider. At DXC, he founded DXC Security Labs, spearheading advancements in cybersecurity research and development.

Craig remains deeply engaged with the private sector, particularly in finance, where he advises major banks and private equity firms on security strategy, architecture, and cyber due diligence. As a member of the Offensive Cyber Working Group's College of Experts, he contributes to research and policy development in offensive cybersecurity. He is also the founding chair of the Geopolitics-Cyber Community of London Experts (GeoCyclone), an organisation bringing together business leaders, academics, and policy makers to share knowledge on national security and digital technologies.
He is currently studying Mandarin (HSK4).

Craig is active on LinkedIn, often posting details of upcoming courses and ongoing research.

Application

Please use the 'Book' button on this page. Alternatively, please contact us to obtain an application form.

You can also register for individual lectures if you do not wish to attend the whole series.

IT requirements

We will be using Zoom for the livestreaming of this lecture series, and you will be able to submit questions via the Zoom interface. Joining instructions will be sent out prior to the start date. We recommend that you join the session at least 10-15 minutes prior to the start time – just as you might arrive a bit early at our lecture theatre for an in-person event.

Please note that this course will not be recorded.

Terms & conditions for applicants and students

Information on financial support