Tell Us About Pimloc, How Did It Come About?
Before launching Pimloc in 2017, I led the OMG Life team that developed and launched the world’s first intelligent wearable camera and its associated technology platform licensing. While building that particular product, I noticed the lack of clear protection for biometric data privacy alongside a need for systems that can automatically classify and surface relevant content from large image and video streams.
To combat these issues, we launched Pimloc – a specialist deep learning service for video security and data privacy applications. The platform offers two key services that allow private businesses and public organisations to protect and responsibly analyse sensitive and personal data in videos and images;
- Pholio: a system for identifying specialist visual content in large collections and streams.
- SecureRedact: a system that automatically anonymises sensitive content in large scale video.
What Problem Are You Trying To Solve?
While regulation has come a long way with regards to privacy, controls on how and where surveillance footage is stored and shared has been left at the wayside. Pimloc’s primary purpose is to prevent biometric data from falling into the wrong hands. Designed for any organisation that captures and stores video, the platform provides automatic anonymisation of sensitive data, without creating the burden of intense manual work for team members.
Companies should be able to keep their premises secure and extract valuable business analytics from video without compromising customer, employee and public privacy in the process. In the UK, Europe and certain US states they need to be able to quickly respond to CCTV based data subject access requests for GDPR and Freedom of Information requests for the same.
Cities and developers should be able to harness live data to better manage facilities, utilities, security and public services whilst also protecting the personal data of their citizens in the process.
People should be able to post images of their friends and family online, or sell cars and clothes via online marketplaces without exposing personal data of their children, number plates and other sensitive identifiers regularly captured in images and video.
What Have You Learnt So Far?
We’ve seen a huge spike in the number of organisations and individuals looking to use our services to redact personal data in CCTV video feeds and for individuals to protect the faces of their loved ones in photographs they publish online.
We’re seeing a growing demand for visual anonymisation services across a wide range of business sectors and consumer use cases. In the early stages this was driven by data privacy legislation requirements but increasingly we are seeing businesses taking more of a responsible and transparent approach to how they collect, store and use personal data. If all the personal and sensitive elements of video are protected by default then the data can be accessed more widely for the benefit of all.
More from Interviews
- A Chat with Katy Wigdahl, CEO at Speech-to-Text Software AI: Speechmatics
- Meet Mario Ciabarra, CEO at Continuous Product Design Platform: Quantum Metric
- A Chat with Octavi Royo, Founder and CEO at Digital Art Platform: Agora
- A Chat with Gustav Schauman, CEO and Founder at Disruptive Golf Sneaker Brand: Goatlane
- Meet Gervase Fay Fernandes, Co-Founder & Commercial Director at Health & Her
- A Chat with Howard Marks, CEO and Co-Founder at Premier Start-Up Investment Platform: StartEngine
- A Chat with David Brown, CEO of Social Enterprise Fintech: Hi
- Meet Christina Kosmowski, CEO at Cloud-Based Monitoring Platform: LogicMonitor
Why Should the Public Care About Surveillance Footage and Privacy?
Video surveillance is everywhere in the UK. London has a CCTV camera for every 14 residents, – due to the sheer number of private businesses and public sector organisations capturing CCTV footage. From banks to restaurants, and shopping malls to tube stations, cameras are here to stay. These surveillance systems are constantly capturing sensitive biometric data, such as faces or genders, and identifiable information such as number plates. Once captured, this data is often stored by private companies in data centres vulnerable to cyberattacks, or shared with third parties for analysis.
If we’re serious about protecting people’s privacy then this is an area policymakers and companies can make a huge difference. By using an automatic video redaction tool, like ours, businesses that are collecting huge amounts of footage can make sure that they keep it protected and secure. People can also have more control over the personal data they are posting online – to social media, dating sites and shopping marketplaces.
What Challenges Have You Overcome?
We are focused on three core areas; machine learning capability, product user experience, and go-to-market.
Unlike many deep learning applications that can manage with moderate accuracy levels, we need to provide high levels of automated performance across a large diversity of real-world CCTV, body-worn camera and smartphone footage. These domains have a very long tail of content types – each camera is mounted in a different position, at varying distances and angles from subjects, they look at different scenes, in varying light levels, in very dynamic environments across a large variety of camera types and quality. It’s very easy to show high percentile performance for AI systems against test data sets but they need to work just as well on real-life, in-the-wild, data in order to protect personal data in video at scale.
We are developing cutting edge technologies but need to package their power into simple tools that non technical users can master very quickly. We are constantly challenging ourselves to improve the usability of our systems, to simplify end user workflows as much as possible to reduce any manual operator time as close to zero as possible. Marrying world-class automation with simple but powerful end user applications means we can cater to the full breadth of video privacy requirements across sectors.
On the go-to-market side, we are investing time to understand the detailed requirements of our users in each sector and how they mesh with local data privacy legislation. We have attracted customers from a wide range of sectors, all with similar user requirements, which we are now focussing on how best to support them through direct sales alongside integrations to third party video platforms.
How Have You Responded During COVID?
The Pimloc team has been incredible. It’s been a tricky year to build and grow a team when we’re all remote – but everyone has risen to the challenge. We were fortunate to raise our Seed round of funding last October which has allowed us to accelerate our technology, products and market approach through the last year.
We were used to remote and hybrid working prior to the pandemic and have been able to continue in that vein since. The biggest challenge has been building culture in a time where it’s been difficult for us all to get together in person; we’ve compensated with socials on Zoom and virtual coffees but nothing is as good as getting everyone together in person.
What Are Your Plans For Growth?
Our focus is now on growth – growth of tech, growth of products and growth of go-to-market. We are focused on building out the Secure Redact business in UK, Europe and the US and starting to develop our future ‘responsible analytics’ platform. We have a roadmap that will allow us to meet the growing levels of demand for our current products, and are accelerating our operations while simultaneously starting to build new capabilities for the future.
Our core mission is to protect even more personal data in new markets and geographies – helping secure sensitive data for us all.