Rogue AI: Can AI Be Hacked and What To Do About It?
Rogue AI: Can AI Be Hacked and What To Do About It?
Part of the CDT Maker Club August Takeover
Can an AI agent be hacked or manipulated into going rogue? Explore real AI security risks and learn how we can defend against them.
Full Luma Description
What happens when an AI agent stops behaving as intended?
As AI becomes more capable and increasingly able to take actions on our behalf securing these systems is becoming one of the biggest emerging challenges in cybersecurity.
Rogue AI is an interactive evening exploring how AI agents can be attacked, manipulated or pushed into unexpected behaviour and, importantly, what we can do to secure them.
Led by Liana Anca Tomescu, CEO & Founder of SonnyLabs.ai, the event will take participants from the fundamentals of AI security through to practical examples of the risks that organisations, developers and everyday users increasingly need to understand.
You’ll hear how vulnerabilities can emerge, see examples of what happens when AI systems are exploited and take part in a hands-on activity using your own laptop or mobile device to experience some of these risks for yourself.
The evening will conclude with a panel discussion exploring the future of AI security, responsible development and what organisations need to start thinking about now.
You do not need to be a cybersecurity expert to attend. The session is designed for beginners and intermediates who want to understand one of the fastest-growing areas of technology and cybersecurity.
What You’ll Explore
During the evening, you will:
Discover how AI agents can be attacked, manipulated or exploited
Understand what people mean when they talk about “rogue AI”
Explore vulnerabilities and emerging cybersecurity risks around AI
Experience practical examples of AI security risks using a laptop or mobile
Learn approaches used to secure AI systems and reduce risk
Hear different perspectives during a live AI security panel
Meet others interested in cybersecurity, AI and emerging technology
Who Is It For?
This event is ideal for:
Students and graduates
Cybersecurity beginners
Developers and technologists
AI enthusiasts
Computing and engineering students
Founders and start-ups
IT and security professionals
Researchers and academics
Anyone curious about how secure AI really is
Level: Beginner to intermediate.
You don't need specialist AI or cybersecurity knowledge to take part.
Event Details
Date: Wednesday 26 August 2026
Time: 6:00pm–9:00pm
Location: Royal Docks Centre for Sustainability, UEL Docklands Campus.
The nearest DLR station is Cyprus, which is directly beside the Docklands Campus. Once on campus, attendees should follow the signs for the RDCS Building/Royal Docks Centre for Sustainability. Knowledge Dock is the building’s former name and will also bring you to the correct location.
The event space is on the ground floor, immediately to the right as you enter the building. There will also be Maker Club signage.
Cost: Free
Level: Beginner & Intermediate
Facilitator: Liana Anca Tomescu
Organisation: SonnyLabs.ai
Bring with you
Participants are encouraged to bring a laptop or smartphone for the interactive workshop element.
Planned Running Order
18:00 - Registration & early networking
18:40 - Welcome and introduction
18:45 - Rogue AI: Security for AI Agents with Liana Anca Tomescu from SonnyLabs.ai
19:30 - Hands-on AI security workshop - Try out your own prompt injections.
19:45- Fireside Chat with Mike Anderson from Covia.ai
20:15 - Networking
21:00 - Event Ends
About the CDT Maker Club August Takeover
The CDT Maker Club August Takeover is a free programme of practical workshops, demonstrations and creative sessions led by members of the UEL and wider Maker Club community.
It provides a welcoming and supportive space to test ideas, develop practical skills and share knowledge across technology, digital making and creative practice.