Ending Violence Against Women and Girls on London buses

The challenge

Context of the challenge

How might we combine new innovations with existing on-bus technology to automatically detect sexual offences on buses reliably, while upholding the highest standards of privacy and compliance?

It’s widely recognised that sexual offences are underreported across society – this is also the case for London’s extensive bus network. For this reason, data from reports to police and/or TfL relying on victim or bystander reporting only provide a partial picture. While buses have CCTV, manual review is reactive and often labour-intensive. Previous trials have shown automated detection is technically feasible but were constrained by privacy requirements and the need for access to real CCTV footage for training.

TfL and Plexal are seeking privacy-preserving detection systems using synthetic data, open-source datasets or pre-trained models that can enable live alerts and/or rapid post-incident investigation without compromising passenger privacy or public trust.

Challenge Areas

We’re inviting solutions in three distinct, but related, areas: 

(1) Synthetic data for model training

Demonstrate effective generation of synthetic data environments that can be used as an alternative to real CCTV footage for training behaviour detection models.  

(2) Motion/behavioural detection  

Technologies (from retail or adjacent sectors for example) that can reliably be adapted to detect sexual offences and other high-harm offences using non-visual sensors or motion analysis, adapted for the bus environment. 

(3) Multi-modal sensor fusion  

Combining non-visual and visual data sources (lidar, audio or vision for example) to enhance detection accuracy, while maintaining GDPR compliance and avoiding facial recognition or continuous audio recording. 

Project scope

This market scouting exercise identifies viable suppliers for potential proof-of-concept trials in FY 2026/27. Successful applications will demonstrate solutions and may be invited to participate in funded pilot deployments. 

Essential Requirements 

Technical capability

• TRL 6-7 (demonstrated in relevant operational environments) 

• Enable detection of at least one type of sexual offending 

• Reliable detection across varying lighting, crowding and camera angles 

Privacy and compliance

• GDPR compliant 

• Grounded in strong data ethics principles 

• No facial recognition; no continuous audio recording 

• No requirement for use of on-bus CCTV footage 

• Transparent and auditable data handling documentation 

 Integration and feasibility 

• Integration with existing on-bus CCTV and iBus systems 

• Minimal/no additional hardware beyond current specifications 

• On-device or edge processing capability  

Out-of-scope 

• Predictive analytics or “pre-crime” detection 

• Individual identification or passenger tracking, including facial recognition technology 

Important Notes on AI and Privacy

TfL is committed to upholding public trust and privacy. This call off is not seeking surveillance solutions. All proposals must demonstrate privacy-preserving approaches. Avoid any perception of intrusive monitoring and be underpinned by strong data ethics principles. 

Application Process

Applications open on Thursday 8th January 2026 and remain open until Monday 19th January 2026. Interested suppliers should submit concise responses (maximum two pages or equivalent), clearly outlining:  

• The challenge area(s) addressed 

• Technical approach and readiness (methodology, TRL, detection capabilities and accuracy) 

• Privacy and compliance measures (training data sources, GDPR measures, privacy safeguards) 

• Integration and deployment plan (infrastructure requirements, hardware needs, timeline, scalability) 
• Evidence of effectiveness (where available) 

Eligibility

Open to all organisation types and sizes: startups, academic institutions, established suppliers, adjacent sector innovators. 

Submit your applications as follows

Email IOS@plexal.com with your completed submission, using the subject: “EVAWG Open Call-Off Application – [Organisation Name])