SIS Technology Challenge

MI6 search for startup solutions

Secret Intelligence Service MI6

Last year, Plexal announced a key partnership with the Secret Intelligence Service (SIS) – also known as MI6 – as part of the SIS Technology Challenge. 

Our mission is achieving the potential of emerging technology through strategic collaboration with the government, industry, startups and academia. In keeping with this, we’re pleased to announce Plexal is once again working with SIS to close the gap between organisations and their problems and solutions, using technology to deliver national security and prosperity. 

SIS wants to discover creative, interesting and novel solutions for a series of technology challenges. Aiming to take advantage of a thriving science and technology ecosystem, SIS recognises the opportunities this presents for national security, with the support of Plexal’s experience creating collaborative ecosystems across the private-public sectors. 

Could your product help protect the UK? Make your mission matter with the SIS Technology Challenge. This opportunity is designed to encourage non-traditional suppliers to work in the national security space, including startups, SMEs, industry, academia and research institutions with solutions that can be applied to one of the challenge areas below.

WHAT’S THE APPLICATION PROCESS?

Our client would like to receive expressions of interest (EOI) and, if appropriate, a solution or approach outline of 1000 words maximum. The client will contact parties they’re interested in following up with.
EOI must be received by Monday 6
th January 2025.  

We’ll be looking to take forward one organisation for each challenge area into a POC this financial year (Pre-April).
If you haven’t heard from us by Monday 3rd February 2025, this means you have not been successful.
 

Please contact apply@plexal.com to register your interest in applying for these challenges and submit your EOI.

THE CHALLENGES

PERSON PORTABLE COMMS

Organisations operating in challenging environments face a critical need for secure, communications, with a low probability of detection, that can function without relying on in-country infrastructure. Common solutions often expose user location and identity through detectable transmissions or metadata analysis, creating significant operational security risks.

SIS is seeking innovative solutions for Low-SWaP (Size, Weight, and Power) communication systems that prevent both physical and digital detection of its officers. The solution must be independent of in-country infrastructures and enable users to maintain secure contact across diverse geographic regions while protecting both location and identity.

Technical solutions must address both physical and digital detection prevention and should support one-way or two-way communication, with device life spans ranging from single use to persistent, multiyear longevity.  

EOI may address either complete communication systems or critical components within a Low-SWaP system. We welcome innovative approaches that balance security, usability and scalability while maintaining operational security in challenging environments.

KEEPING AHEAD OF MODEL DRIFT 

SIS uses Artificial Intelligence (AI) and Machine Learning (ML) to enhance its efficiency and capability. Examples of this are pattern recognition from financial records, identifying and countering cyber-attacks, large scale simulation or digital twinning and large language models. To maximise the effectiveness of AI deployments, SIS must monitor ML model performance and retrain systems at appropriate times. Detecting model drift, a common cause of performance degradation, is a critical part of this.  

Model drift emerges from shifts in the data distribution or relationships between inputs and targets over time. As new topics emerge, vocabulary evolves and global landscapes transform, both concept drift and data drift are common. It’s necessary for SIS to detect and respond to challenges which arise from model drift.  

There are a range of sub-problems that contribute to this challenge, which could include the increased complexity of unstructured data, multi-modal drift detection, reactive vs. proactive detection, drift threshold definition, the need to limit model output storage and human intervention, granular segment detection, explainability and interpretability of drift, the ability to operate on premises and accounting for data quality and bias.  

EOI may address either complete drift solutions or look to target one of the sub-problems outlined above. Solutions should seek to balance detection accuracy and performance with supportability, security and usability.  

INTERESTED TO KNOW MORE?