Venturing into AI and Data Science
As AI rapidly integrates into our daily lives, here are some insights and use cases of the technology to improve critical services
In just a few months, ChatGPT has spurred the rapid, commercial growth of AI tools – and a louder cry for internet safety, job security, combating biases and discrimination, and other concerns of the risks of AI.
In 2021, the UNICEF Venture Fund (the Fund) saw the potential of the then-emerging technology to positively impact learning, health, and psycho-social support opportunities. It then launched a call for applications for startups developing Open Source, AI and machine learning solutions that accelerate learning outcomes, generate data to forecast health and health care needs, and provide access to online services at lower costs, even with low connectivity settings.
And, with children and their families turning more to digital solutions, they find a world of opportunity that may also increase children’s exposure to online risks and harms. The same technologies must be explored to efficiently prevent and respond to the threats of this new online environment for children.
In response to the increased threats online, and in partnership with the Global Partnership to End Violence Against Children, the Venture Fund invested in two new startups developing Open Source, AI-powered solutions to address digital risks to children. The Fund invested in 11 solutions through two cohorts – Child Online Safety and AI and Data Science-driven systems solutions. These startups have spent the last 12-18 months building, testing, and validating their solutions.
As the startups complete their investment period with notable results in improving services for children and their communities, such as enabling access to improved healthcare for 6107 users, UNICEF reiterates its call for frontier tech to become tech for good.
In the June 2023 Venturing into AI and Data Science: Safe Solutions for Health, Education, and Climate event, the work of these startups was presented alongside timely discussions of the opportunities and challenges of AI with renowned experts in the space. These conversations, moderated by Startups Without Borders founder Valentina Primo, featured Thomas Anglero, Nordic Chief Technology and Innovation Officer, Cognizant, and Jon Krohn, Co-Founder and Chief Data Scientist, Nebula.io and host of leading data science podcast, SuperDataScience.
The event kicked off with Thomas Anglero sharing his vision beyond this “Internet Explorer” moment of AI and the immense opportunity for the technology to better our world.
Anglero’s remarks were followed by presentations by six startups working to strengthen safety for children online and improve access to education. Since the beginning of the investment period, 2M+ youth have been provided with additional skills to support social, emotional, and mental wellbeing. Additionally, 184,715 users have had improved access to educational tools.
After hearing use cases of how AI has a huge opportunity for good, Jon Krohn expressed how AI’s technological pace and progress can play a critical role in solving challenges facing people.
However, Krohn cautioned that the unprecedented speed of AI also leads to risks like “AI-driven news feeds or bias in training AI models that reinforce stereotypes.”
The 11 startups, all from emerging markets and all leveraging data science and AI in their solutions, are well on their way to solving real-world problems and ensuring children can be safe online. We welcome collaborators from all backgrounds to support these teams toward unlocking the potential of AI to change lives.
An overview of the startups and opportunities for engagement are presented below.
Tilli (Sri Lanka): Game-based, AI-powered social-emotional learning tool
- Developed an AI-powered social-emotional learning tool with 80% completion rate
- Over 90% of learners reporting successful integration of learnings with everyday problem solving
- 5,000+ learners have benefited from the solution to date
- Collaboration with Save the Children Sri Lanka reached over 3,000 families
Needs for further growth and opportunities to scale:
- Pre-seed and EdTech/mental health funders to join/lead their pre-seed round in raising $1.5M for three key areas:
- Expanding Tilli’s Gen-AI powered learning suite to build, map, and measure 8 essential cognitive skills from self awareness to emotional resilience.
- Build engineering and product capacity especially in the fields of machine learning and data and learning sciences.
- Scaling to 100 elementary schools
- Connections with elementary schools, INGOs, and education ministries to support their short-term goal of scaling to 100 schools.
Talk2U (Brazil): Digital therapeutics platform to improve social, emotional, and mental wellbeing
- Launched the “Como Se Eu Fosse” WhatsApp chatbot to combat online hate speech, reaching over 500 participants
- 85% of users found the experience to be valuable to learn about hate speech and deal with its impact.
- Reached over 2 million beneficiaries to date across 8 conversational journeys (example topics include, Mental Health, HIV, and more)
Needs for further growth and opportunities to scale:
- Funding to support app development, B2C business model development, and efficacy trials.
- Individuals and organizations collaborating on the SDGs that can sponsor new content to expand impact, running cultural and language adaptations to existing journeys within new regions, and researching new target audiences to map new journeys needed.
Bookbot (Indonesia): Gamified app with real-time, on-device speech recognition technology
- Launched gamified app, with a library of 2,500 books in Bahasa and English with incentives and reporting capabilities
- Collaborated with INOVASI and Indonesia Ministry of Education and Culture to provide professional development for 415 teachers with its digital literacy program
- 32,000 app downloads
- Recorded increase in average reading time per child and doubled average reading proficiency of children within one year
Needs for further growth and opportunities to scale:
- Partnerships with governments, IGOs, and NGOS to improve literacy in Indonesia and beyond. Solution can be adapted for other language use cases.
Om3ga (Serbia): Deep learning speech-to-text solution for Slavic languages
- Built Daktilograf solution offering fast and accurate transcription of unlimited numbers of words via online platform, mobile app, and on-premises solution. Product has been adapted for offline AI text-to-speech and speech-to-text for South Slavic languages. This is the first and largest open public TTS and STT dataset for machine learning in the region.
- Secured media partnerships, including with Nova Ekonomija.
Needs for further growth and opportunities to scale:
- Introductions to systems integrators (B2G, B2B, and B2C) in education, media/entertainment, and other relevant sectors.
AfriLearn (Nigeria): Accessible, adaptive, and affordable learning for Africans everywhere
- Developed an AI-powered recommender systems to deliver personalized learning experiences to users
- Integrating gamification to drive growth while making learning fun and engaging for students
- Serving 20,000+ direct users and 400,000+ indirect beneficiaries, including parents, teachers, and schools
Needs for further growth and opportunities to scale:
- Partnerships to scale reach and impact across and beyond Nigeria to West Africa.
Jobzi (Brazil): AI and data intelligence to predict the impact of school connectivity on labor and education outcomes
- Released data and insights, APIs, and SAAS platform accessible at unicef.jobzi.com
- 210,000 B2C users and 50 corporate clients
Needs for further growth and opportunities to scale:
- Fundraising $1M to extend business to other languages and parts of the world.
Eyebou (UAE): AI for eye screening tests for children to detect vision disorders
- Screened over 10,000 children, giving them access to eye care
- Funded by SOS Children’s Villages UK to pay for lenses for children in need
- Achieved high ease-of-use rating from users in feedback sessions
Needs for further growth and opportunities to scale:
- Institutional partnerships with clinics and health NGOs to license their technology to screen over 20,000 children.
Portal Telemedicina (Brazil): population health management platform using AI to detect anomalies and prioritize emergencies
- Launched 3 modules into platform: Data, AI, and Dashboard modules, integrating multiple data sources
- Launched user-friendly dashboard interface
- Partnered with City of Taruma to integrate Smart Child Development Platform, with a focus on vaccination data
Needs for further growth and opportunities to scale:
- Expertise in navigating regulatory and legal frameworks in different countries
- Potential partners and customers
- A Series A fundraising round to support the international expansion of the project.
Cirrolytix (Philippines): Dengue prediction and epidemic management platform using climate and health data
- Project AEDES work featured on Data.org as an innovative approach to early detection and exploration of disease risks
- Certified as a digital public good (DPG), signifying its value in addressing global challenges
- Collaborating with the Philippine Pediatric Society – Southwestern Mindanao Chapter to implement clinical interventions based on prediction models
Needs for further growth and opportunities to scale:
- Institutional partnerships for co-designing multimodal, multidimensional dashboard and analytics for other types of risks (I.e. climate, disaster, political risk, ESG).
AQAI (India): Machine learning powered predictive model for monitoring air quality
- Developed the 1km x 1km open source PM2.5 (Particulate Matter) API
- Collaborated with OpenAQ on global Particulate Matter data, resulting in a high fidelity Air Quality Model
- Collaboration with UWaterloo and UNICEF Mongolia, creating models for health effect measurement in Mongolia
Needs for further growth and opportunities to scale:
- Partnerships and pilots to integrate Air Quality product and expertise with a strong emphasis on collaborating with healthcare providers.
Neural Labs Africa (Kenya): AI to accelerate medical imaging diagnosis for respiratory diseases
- Onboarded 5 hospitals for clinical trials
- Screened over 5,000 images in Kenya and Senegal
- Accepted to Vilgro Africa incubator for emerging healthcare business in Africa.
Needs for further growth and opportunities to scale:
- Partners and access to data to scale up infrastructure and R&D.
