🎉 JOGL is soon launching a new version. All the users of the v1 will be migrated to the new version. In the time being, we do not allow the creation of new users on this platform.
AccuroLab - Factcheck Covid-19 Information banner
Project
9
Members

Status:
Active/Ongoing
Linked to group(s)/challenge(s):

AccuroLab - Factcheck Covid-19 Information

About reviewed project
A platform to fact check Covid-19 information received on social medias and provide users with shareable verified information.

1.0 Introduction

AccuroLab solution was awarded at the MIT COVID-19 hackathon of May 2020. The goal of our team is to combat Covid-19 misinformation where fake news are the most shared. We are putting in place a solution that will allow users of poor communities to validate the accuracy of the Covid-19 information they received on social media through a fact checking process against trusted sources.

 <> Our project team operates on the #accurolab channel of the OpenCovid19 JOGL slack.

 <> The project prototype v01 demo, presented to the MIT panel, is available here


1.1 Problem and Background Summary

Fact-checking information and accessing verified sources was a crucial part of being well-informed. But the COVID-19 crisis has made this step critical to keeping vulnerable African users and communities safe and healthy.


Fake news spread faster and more easily than this virus 

Tedros A.G, WHO Director General, Feb 2020 


Here are some examples of Fake news messages, We (Team members) have received from our families living in Africa and Abroad:

Here is the persona our project is working hard to reach during the Covid-19 Pandemic:


Problem statement: Smartphone users between the ages of 35 and 55 from developing countries in Africa need a way to verify information that they receive and share concerning the prevention, management, and statistics of the COVID-19 pandemic.

In 2018, there were 340 millions smartphone users in West & Sub-Saharan Africa. This figure is expected to reach 670 millions users by 2025.


1.2 Solution summary in simple terms

AccuroLab is a mobile-based messaging system with 3 goals:

<> Raise the awareness of cell/smartphone users over the dangers of Covid-19 infodemic

<> Provide cell/smartphone users with an easy-to-use fact check tool for Covid-19 information they receive 

<> Provide cell/smartphone users with an information extraction channel through which they can pull or subscribe to the latest Covid-19 updates provided by a selected list of Trusted Health Organizations (THO)


AccuroLab acts as an intermediary channel between cellphone users and THO sources: when a user subscribe to the AccuroLab service, AccuroLab’s fact check functionality allows the users to assess the accuracy of their incoming texts related to the coronavirus by using a key-words and key-phrases search engine and provides the user with succinct, easy-to-understand bullet-point card information compiled from a regularly updated list of THO database.


Ongoing prototype developments focus on allowing AccuroLab to fact check Covid-19 information related to 4 thematics:

<> Causes & Transmission

<> Symptoms

<> Treatments & Prevention 

<> Statistics (Local, Regional & Worldwide)


1.3 Solution summary in technical terms

To reduce user friction and circumvent the moratorium on new coronavirus app submissions to the Apple App and Google Play Stores, the design of AccuroLab has evolved from a mobile application to a cloud-based messaging application that users can interface with by sending Whatsapp or SMS messages using the Twilio's messaging APIs. With this design, our users can register easily, quickly and without any download requirements. 


AccuroLab ‘s core functionality (Fact Check tool) works as described below:

<> Front-end: Designed to integrate with both Whatsapp & SMS messages, AccuroLab’s fact check tool utilizes an NLP (Natural Language Processing) model to decipher the piece of text sent by a user, and check the text against its categorized database of trusted sources. It uses the process of related entities and embedded words to identify related Covid-19 information tokens. 

<> Back-end: After having received an incoming text from the user, the information is processed by the backend server. In a matter of seconds, AccuroLab’s backend server launches a matching process that queries our pre-built Coronavirus Knowledge Base (CKB). The CKB is a database of categorized Coronavirus information snippets from THO websites. Finally, the back-end sends back relevant information from matched categories to the end-user. As new information arrives from the list of AccuroLab THOs, AccuroLab’s CKB is able to stream new information and update its database of verified content.


This process will allow AccuroLab to search the data sets that have been fed to its database, categorize the user queries, and extract the most relevant information from the THO’s websites. Upon receiving the results, the user can then assess the accuracy of the text he/she received against the information AccuroLab helped him/her obtain from the trusted sources.


1.4 Originality

The originality of AccuroLab is based upon 3 points:

<> AccuroLab is not a chatbot: In contrast to existing COVID-19 Chatbots that provide a limited number of request options from the user and deliver long and generic information, one of AccuroLab’s key added value lies in its ability to fact check incoming text messages by mapping their content against a list of verified sources and delivering content-related and verified information cards to the end user.


<> AccuroLab is an information dissemination tool designed to act as intermediary between THOs and vulnerable communities: Due to language diversity, bandwidth, network coverage and prepaid data challenges (especially in Africa), our market analyses indicate that AccuroLab’s success rate in the African market lies in its ability to:

<> Integrate seamlessly with both Whatsapp & SMS platforms which are the most used for sharing text-based content in the majority of African regions

<> To deliver verified, concise and content-related translatable information to users (based on their pull requests or on the dissemination strategies of local or regional THOs)


<> Beyond its Covid-19 scope, AccuroLab aims at raising the awareness of its users on the dangers of health-related misinformation. Thereby, as a tool, it provides Governmental authorities & Trusted Health Organizations with a channel to disseminate written & graphical content to help users of developing countries to build up new critical thinking behaviors when analyzing news.


1.5 State of advancement of the project

During the project ideation phase at MIT COVID Hack (May 1-3), a first mobile application prototype (v0.1) was developed. Following more advanced market analyses and workshops with our mentors and potential users on the ground, the team moved quickly to launch 3 streams of iterative action on May, 4:

<> Refine both our problem and solution statements

<> Design prototype v0.2 architecture and launch a back-end engine development of AccuroLab’s fact check tool

<> Define, collect and structure AccuroLab’s Coronavirus Knowledge Base (CKB).


1.6 Project Timeline

The goal of the AccuroLab Project team is to deliver a running fact check tool prototype (v0.2) on June, 3. The team plans to use our v0.2 prototype to launch a first wave of béta testing on a limited scope of pilot users being identified (15-25 users).

Our next steps are:

<> Finalizing the architecture design of prototype v0.2 : 1-2 working days

<> CKB Data structuring & information card creations: ~3 working days

<> Twilio API Back-end services configuration: ~3 working days

<> Fact Check tool Development, Configuration & Unitary testing:~5 working days

<> Demo creation: 1 working day


2.0 Project Implementation

2.1 Our solution

Our goal is to provide smartphone users of vulnerable communities with an ergonomic and easy-to-Get & Use solution that will cover 2 key functions presented here below. The solution has the potential to be used to limit the damages of misinformation during any large health pandemic, just like the one with are currently facing with the COVID-19 Pandemic.

AccuroLab is a:

<> Fact-checking Tool: With AccuroLab, Social Media Users factcheck the information they receive. AccuroLab assesses the accuracy of the scanned content by crossing it with a base of reliable data from trusted sources, and direct Users towards reliable sources.

<> Crisis-compass Tool: As a dissemination tool, AccuroLab’s Users receive verified information from trusted sources. Additionnally, Government & Health agencies push out verified information and guidelines to local communities via AccuroLab to help them navigate safely with certified information during the pandemic crisis.


The v01 prototype demo of AccuroLab is available here

2.2 State of the project & Timeline

Since the MIT Hackathon of May 1-3, the team has been moving quickly to move our product towards deployment. On the market research side, we first conducted larger user interviews with Africans in our user demographic and collected quantitative and qualitative data we will be using to refine our product.

We also conducted industry and domain research to understand what is out there in terms of combating misinformation and found a few products that were not similar to our value proposition. Our initial prototype (Prototype v.01) is being refined to accommodate identified frictions and improvements (Prototype v.02 being implemented).


A/ Ideation phase: May 1-3, 2020

The AccuroLab project has started on May 1 during a Hackathon organized by the MIT: Africa takes on COVID-19. This 48-hours challenge during which +1500 individuals from +100 countries competed to tackle Covid-19 related challenged in Africa has been our project playfied ideation. 

During the Weekend between May 1-3, the project team:

<> Has decided to focus on combating the Covid-19 infodemic issue

<> Defined an initial problem statement

<> Gathered the pain points of some end users on the ground in order to refine the project scope


We conducted a short but deep dive market analysis by interviewing +20 african users based in various African countries of North, West, Central and Southern Africa regions.

<> This phase was critical in identifying through which medias / channels the infodemic where spreading the fastest (Devices: Smartphones VS Cell Phones VS Tablets VS Computers; Social media: Whatsapp VS Facebook VS text messages, etc..

<> This phase also allowed us to characterize our persona: his/her behavior, paint points and needs as an end-user.

<> Finally this phase was critical in identifying the first frictions we should be able to address in our prototype v0.


The team also conducted a market research on a more global scope to identify ongoing initiatives aiming at combating Covid-19 infodemic, their current limitations and the potential key actors / partners to get in touch and whose experiences can be leveraged to design a solution more customized for African communities.


On May 3, at the end of the MIT Hackathon, the key deliverables of the team were a deck and a prototype:

<> The deck formalized where and how the AccuroLab's team intended to combat misinformation: problem statement, solution statement, solutions functionalities and targeted market

<> The prototype demo showcased the key functionalities of the solution and an assessment of its feasibility.


B/ Refinement phase: Since May 4, 2020

Since the end of the MIT-Hackathon, the team has been moving quickly on 3 aspects:

<> Problem statement refinement & Go-To market strategy definition: We have conducted additional user interviews with Africans in our user demographic and collected quantitative and qualitative data we will be using to refine our product and how to make it easily usable for our target End user. We also conducted more industry and domain research to understand what is out there in terms of combating misinformation and found a few products that were not similar to our value proposition.

<> Prototype enhancement and back-end engine development: Our initial prototype (Prototype v.01) is being refined to accommodate improvements. Architecture Design has evolved, data structuration is ongoing and Prototype v.02. is being built.

<> Implementation plan definition & financial evaluation: We have also put together an implementation and financial plan to ensure that we are covering our basis in terms of OPEX and CAPEX, and to keep us on track to deploy the product as soon as possible.


Our next steps are:

<> Finalizing the architecture design of prototype v0.2 : 1-2 working days

<> CKB Data structuring & information card creations: ~3 working days

<> Twilio API Back-end services configuration: ~3 working days

<> Fact Check tool Development, Configuration & Unitary testing:~5 working days

<> Demo creation: 1 working day



2.3 Methodology

AccuroLab is a mobile-based messaging system with 3 main objectives:

<> Raise the awareness of cell/smartphone users over the dangers of Covid-19 infodemic

<> Provide cell/smartphone users with an easy-to-use fact check tool for Covid-19 information they receive 

<> Provide cell/smartphone users with an information extraction channel through which they can pull or subscribe to the latest Covid-19 updates provided by a selected list of Trusted Health Organizations (THO)


Our first 2-week development sprint focuses on the fact check prototype v0.2. Once we have a running prototype, we want to open it up to end-users for feedback. This will allow us to identify initial user-experience constraints and give enough data to start improving our product output, solution value-added, and users experience.


At the end of this development cycle, AccuroLab will act as an intermediary channel between cell phone users and trusted health organizations (THO) sources on one or many of COVID-19 information categories presented below:

<> Causes & Transmission

<> Symptoms

<> Treatments & Prevention 

<> Statistics (Local, Regional & Worldwide)


Our global execution plan is designed on a 10-Weeks time frame as presented below: 

<> 4 weeks sprints of development (this phase includes the 2 weeks presented above)

<> 3 weeks of system scaling & optimization

<> 1 week of pilot phase

<> 2 weeks of Go-Live preparation for a given geographic area (being identified)


The global planning below specifies our execution planning for the next 10 weeks.

2.4 Expected results

Considering both AccuroLab’s 3 goals and the traction factors mentioned in the dedicated section of this document (see section 5.0), we will measure our success through 3 indicators during the next 3 months:

<> The number of end users of developing communities signing up to our platform 

<> The number of fact checking requests received

<> The number of infodemic self-awareness contents (written or graphical) pulled through our platform


Following the first béta testing planned in 2 weeks (15-25 users), we expect to get valuable insights from the field in order 

<> To enhance and stabilize our solution for a pilot phase 6 to 8 weeks post-béta testing phase (50+ users)

<> To target and attract more visibility through trusted local figues (Social media influencers, Governmental authorities, Communities leaders of the area being identified)

<> To invite potential THOs willing to leverage our platform to push materials to fight against the COVID-19 Infodemic.


We finally expect the contributors of the JOGL community to take part, share and leverage the finding of the project to both empower our initiative but also theirs.


3.0 Safety, quality assurance and regulation

AccuroLab acts as an intermediary channel between Cell phone users and THO sources: Therefore, AccuroLab will be neither the source nor owner of the information disseminated. Thereby, each communication coming from the AccuroLab system will include both the reference of the Trusted Health Organization which is the source of the shared information and the date at which the data has been delivered.

In addition, in perspective of the GDPR guidelines, no personal data is / will be stored by the AccuroLab system.


4.0 Impact, issues and risks

Although we expect this project to have impact in multiple dimensions, we pay careful attention to its potential risks and friction points.

In term of impact:

<> AccuroLab has the potential to increase the self-awareness of millions of inhabitants of poor communities about the dangers of the misinformation (On the Covid-19 specifically and more globally on the healthcare system)

<> It also has the potential to give the end user an easy and reliable way to fact check any text-based information related to the Covid-19 pandemic 


In term of potential concerns and risks we closely monitor:

<> A strong emphasis is being and will continue to be put on the identification of Trusted Health Sources (THO) and the monitoring of the lifecycle of the information provided by these THO. This point is all the more critical because of the newness of the COVID-19 disease and the frequency of its statistical and health-related updates 

<> The integration of AccuroLab’s technological stacks with the Cell/Smartphones platforms being the most used in developing communities.

<> A final concern identified has been around the translation of the Covid-19 rich card into local languages. Thanks to MIT, the project team has already reached out to a translation company specializing in +100 African languages in order to prepare contents for the top communities targeted in the early RUN phase of AccuroLab.


These concerns and risks are already in the radar of our workshops with the mentors and contributors who joined us and are supporting us in various fields (Legal, Health, Machine Learning, AI, UX, etc.).


5.0 Traction

The team has developed a Taylor-fitted Go-To-Market strategy that takes into account African communities specificities. Our goal is to deliver AccuroLab in a way such as the solution will be easy to get, easy to use and customized enough to be used efficiently in the developing areas targeted.

Our team has been able to leverage its network to put out information about our success at the hackathon. We are also working with some Universities such as HEC Paris and City College of New York to issue two newsletters highlighting our solution and accomplishment. 


Our traction strategy is based upon four components:

<> One, build our audience by leveraging existing networks (i.e. LinkedIn, Facebook, Twitter, Instagram, WhatsApp) to put the word out, and engage our audience on multiple platforms through short stories, blogs, tweets, podcasts and short videos to generate interest.

<> Two, emphasize product shareability taking into account African communities specificities (Prepaid data contract, Bandwidth & Network coverage, etc.) and generate a Link/USSD Code that makes the solution easily shareable among users on both Smartphones and Cellphones.

<> Three, optimize SEO by creating a website for the solution and using the website home page as a central tag page for all generated content (podcasts, videos, blogs, tweets...etc.) to amplify online traffic.

<> Four, we intend to release, measure, and iterate our solution by collecting feedback from our early app adopters, and using the input to measure for product viability and user experience. We will do this by converting user feedback into applicable product improvements and testing new features through controlled A&B tests to measure for real traction.


The team has also put together an implementation and financial plan to ensure that we are covering our basis in terms of OPEX and CAPEX, and to keep us on track to deploy the product as soon as possible. For a more detailed look at the implementation and financial plan, please reach out the project core team.

As part of JOGL grant application, a section of the financial plan is presented in the funding section here below.


6.0 Team experience

6.1 The Founding Team

AccuroLab Project was founded by:


<> Steve Tchuenté: Steve is an MBA Candidate at HEC Paris & Yale SOM with +7 years of experience in Technology Strategy Consulting in Europe, Africa & North America. He has a strong expertise in advising C-Level executives of various Multinational corporations in defining and implementing their strategic business and technology programs. Learn more about Steve on LinkedIn (Paris, France)


<> Tarik Fathallah: Tarik is a Graduate student and Social Innovation Fellow at the Colin Powell School for Civic and Global Leadership. With 5+ years of experience in his field, he works with a number of community-based organizations to design and evaluate their social impact programs and leverage technology and communication platforms to optimize community engagement & outreach. Learn more about Tarik on LinkedIn (New-York City, USA)


<> Marilyn Osei (Alumna)Marilyn is an Austin-based Product Designer dedicated to designing human-centered experiences. She has extensive experience (+7 years) in finance and business strategy consulting, having worked in banking, investment management and strategy consulting. Today, she leverages her experiences to help design the next generation of digital solutions. Learn more about Marilyn on LinkedIn (Austin, USA)


6.2 Key contributors

Recently, AccuroLab Project onboarded 2 new team members.


<> Nana Kwame Owusu is a Software Engineer with 5+ years of experience in building full stack applications and intelligent systems (AI, ML, NLP) used by millions. He is currently working on his Master's degree in Computer Science from the University of Texas at Dallas. He is passionate about applying machine learning techniques to solving today's problems. Learn more about Nana on LinkedIn (Austin, USA).


<> Sean Anggani is a software engineering professional who focuses on the why and not just the how to build software. He holds a Bachelor in Computer Science from the University of Michigan and is a 3 years experienced full stack developer. Sean is willing to dedicate his expertise to combat the Covid-19 infodemic. Learn more about Sean on LinkedIn (Michigan, USA).


Additionally, AccuroLab core team has a large group of motivated mentors and contributors who are involved with providing their technical and functional expertise around machine learning, governmental health policies, data analyses, data visualization, software engineering etc. A complete list of contributors is shareable on request. 

In term of diversity of expertise, We have also been able to attract two additional mentors: Daniel Lee, a Data Scientist and Brian O’Neill a designer (designing for analytics). A new software developer specializing in mobile app, machine learning, artificial intelligence and NLP is also being hired. 


7.0 Funding and Costs

Following the MIT COVID-19 Africa Hack, the project team has submitted a continuation survey to the MIT administration to confirm our willingness to move ahead on the project and to apply to proposed funding and Grants Programs. Until now, the continuation surveys are still being analyzed by MIT and the project has not received any grant / funding.

So far, our project has been mainly operated on the basis of the conviction, commitment, volunteering and availability of its founding members and the new team members. The JOGL Grant will be critical for us in order to cover the following coming expenses that are essential in moving ahead toward a beta testing phase of AccuroLab:

<> Infrastructure costs (ml.t3.medium instances) on AWS + Storage costs : 500 EUR

<> SME Support on Fact Check tool development (Architecture design validation, Code review, whatsapp & ussd integration, unitary testing): 1200 EUR

<> Integration of Twilio API back-end services configuration: 500 EUR

<> CKB Data structuring, information card creations & 1000 words translation package: 400 EUR

<> Béta-testing Campaign: Preparation, Key users involvement, Communication & Advertising: 50 EUR


We would benefit tremendously from the JOGL Grant, especially if we are awarded the full 2650 EUR, as it would allow us to implement our action plan and fund the costs mentioned above within the next weeks. With strict budgeting, these funds could help extend our expense runway until at least June,10. Additional smaller funding amounts (or volunteer contributions!) would enable us to get individual bits of this work done as well and help us deploy our solution faster in the communities where it is critically needed.



You can reach out to the Project Team by email at accurolab@gmail.com

Additional information
  • Short Name: #AccuroLab
  • Created on: May 12, 2020
  • Last update: July 12, 2021
  • Grant information: Received €2,650.00€ from the OpenCOVID19 Grant Round 3 on Invalid Date
Keywords
Machine learning
Artificial intelligence
Social impact
African market
Ussd code
+ 3
1No Poverty
3Good Health and Well-being
4Quality Education
10Reduced Inequalities