recsyslabs - privacy preserving personalisation for publishers

recsyslabs is a start-up company using artificial intelligence (AI) to make it easy for publishers to deliver tailored newsletters based on readers’ interests, whilst also preserving their privacy and making it easy for them to see why specific content was recommended. 

The recsyslabs technology uses AI to automatically segment readers, based on their interests, and generates one newsletter for each segment by selecting the most relevant content for each subscriber base.

The founders of recsyslabs are Dr Ernesto Diaz-Aviles, Dr Claudia Orellana-Rodriguez and Dr Igor Brigadir, and are affiliated to UCD School of Computer Science. recsyslabs received funding from the Enterprise Ireland Commercialization Fund and spun-off in July 2021.

recsyslabs was one of five emerging startup participants in the annual Nova UCD venture launch accelerator programme in November 2020. The VentureLaunch Accelerator aims to equip UCD researchers with the knowledge, skills and understanding to create a commercial venture with a viable business plan, out of their research.

The recsyslabs team presented their research at RecSys '21: Fifteenth ACM Conference on Recommender Systems in September 2021. Read the text in ACM Digital Library here 

Ernesto from recyslabs explains the background to their technology and how the solution they have developed (NU:BRIEF) helps publishers to tailor and target content better.

“People around the world are becoming uneasy with how much personal data they are asked to give away. They are becoming concerned about how the data they give to web sites and apps is shared with, sold to, or used by 3rd parties, and the lack of transparency about how this works.

Regulators are enforcing stronger data protection and privacy laws such as General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) to protect the public from over-intrusive data collection. Oceans of personal data are vulnerable to loss and misuse, with headline data breaches and hacks heightening privacy concerns.

The use of third-party cookie (tracking) technology is being phased out. However, for many years, cookies have been an important ingredient in the business model for adtech companies, e-commerce businesses and advertising-funded publishers. However intrusive ads and click-bait are not a viable business model any longer. Neo-media publishers, thinking more as a tech company than the old-media counterparts, are extremely focused on engaging with their audience in a respectful long lasting relationship, which leads to more sustainable reader-supported and membership business models. 

recsyslabs is initially focused on the needs of publishers who want to renew and improve their relationship with their readers, regain their trust and loyalty and cut their reliance on intrusive tracking. Responding to privacy concerns and declining ad revenues, they want to build alternative revenue streams around a closer relationship with readers, and subscriptions and memberships. 

Readers want relevant content without the feeling that their privacy is being compromised.  Writers want their words to reach a large and interested audience.

recsyslabs has built an AI product for publishers called NU:BRIEF that personalizes content and delivers delightful recommendations based on user interests without the need for intrusive and excessive reliance on personal data. 

Importantly, it does not impose a big technical burden on publishers to deploy. It does not rely on them gathering, managing and holding securely large pools of personal data, in compliance with data protection law.

Stakeholders (users, customers, readers, advertisers, regulators, the public) can see and understand how the system works and how it came to its decisions, which not only reduces bias and bugs, but also builds trust.

It balances privacy with personalization, giving power and control back to the reader, whilst also delivering the timely, relevant and useful content they want.”

Published 28.09.21