Revolutionizing Media Accuracy Success: How our AI-Powered Tools Are Transforming Media Representation and Fighting Misinformation in Nepal
15 min readUpdated May 13, 2024
Omdena
Omdena Nepal Country Lead Utkhasha Khanal, at AI Conference on the Future of Work in Kathmandu, Nepal
We are thrilled to announce the successful completion of the first milestone in our collaborative project with The Asia Foundation in Nepal and USAID Civil Society and Media (CSM) with outstanding achievements!
It is a part of a pioneering initiative aimed at revolutionizing media representation for women, youth, and marginalized groups through Artificial Intelligence means.
This transformative AI Tool, completed in just six months, marks the beginning of a promising five-year collaboration that promises to change the media landscape in profound ways.
The scale of what we’ve accomplished is not just in the tool itself but in how we built it—by empowering the local community with AI and machine learning skills. This approach doesn’t just solve an issue; it builds the local capacity to keep addressing these challenges moving forward.
Find out how we did it!
What makes our project stand out?
In just half a year, we’ve set a new benchmark for how technology can facilitate societal change, using cutting-edge AI techniques such as natural language processing to analyze and rate media content for its diversity and reliability.
Stay tuned as we continue this exciting journey, pushing the boundaries of AI and community empowerment to foster a more inclusive media environment. Join us in following this impactful story, and discover how technology and collaboration can create a brighter, more inclusive future.
The Partnership
About The Asia Foundation
The Asia Foundation is a nonprofit international development organization committed to improving lives across a dynamic and developing Asia. Informed by six decades of experience and deep local expertise, our work across the region addresses five overarching goals—strengthen governance, empower women, expand economic opportunity, increase environmental resilience, and promote international cooperation.
About Omdena
We are a platform helping mission-driven organizations and startups build impactful AI solutions through global collaboration and empowering AI Engineers from all over the world to become changemakers. To date, over 20,000+ AI Engineers from 120+ countries have been part of Omdena, building impact-driven solutions for 250+ organizations. Omdena also has local chapters in 195+ locations across 80+ countries.
The Problem
Why Is This Representation in the Media Important?
Media shapes perceptions and narratives in society, and ensuring diverse and inclusive representation is crucial for a fair portrayal of all community members. Through this project, we’ve introduced a Comprehensive Media Analysis Tool, specifically designed to identify and improve how these groups are represented in media across two major provinces in Nepal.
Global Challenge of Biased Media Representation
Across the globe, the representation of women, youth, and marginalized groups in the media is often fraught with biases. These groups are frequently portrayed negatively or ignored altogether, leading to societal misunderstandings and perpetuating inequality. Everyone deserves to have their story told accurately, contributing to a more comprehensive understanding and respect among all people.
Why is it a Problem in Nepal?
In Nepal, the issue of biased media representation is particularly acute due to the country’s diverse cultural and linguistic landscape. This diversity adds layers of complexity to how stories are told and understood. Marginalized communities often find their narratives overshadowed or misrepresented, contributing to their ongoing exclusion from mainstream society. The rapid proliferation of media platforms, including social media, radio, and newspapers, complicates the ability to monitor and rectify these biases effectively.
The Urgent Need for Advanced Monitoring Tools
The sheer volume of content and the speed at which it is disseminated make manual analysis impractical, necessitating advanced technological solutions to comprehensively analyze and address these disparities. Additionally, there is a significant gap in the tools available for analyzing media content that can scrape and interpret data from various sources accurately. Without the appropriate tools, media professionals are ill-equipped to identify the extent of bias, let alone address it, resulting in the perpetuation of stereotypes and a lack of diversity in media representation.
Imagine a future where every story is told, and every voice is heard.
The Goal
The primary goal of this project, in partnership with The Asia Foundation and supported by the USAID Civil Society and Media initiative, was to develop a sustainable and impactful tool for enhancing media representation in Nepal, starting with Madhesh and Lumbini Provinces
By implementing a sophisticated AI model, we aimed to streamline the traditionally manual and time-intensive process of media analysis. This innovative approach was to allow efficiently parse vast amounts of data from online news, social media, and radio broadcasts in the Nepali language. Our objective was to foster a more accurate and diverse media landscape, significantly reducing biases and improving the portrayal of all communities within these key regions.
The Background
Outdated Tools and the Impact on Media Professionals
The current media monitoring tools in Nepal are outdated and incapable of effectively processing the vast amount of data generated daily across various media platforms. This technical deficiency prevents media professionals from addressing biases in their content accurately.
Partnership with The Asia Foundation
Our partnership with The Asia Foundation in Nepal aims to tackle these challenges by leveraging advanced technologies such as AI and natural language processing. These technologies are crucial for developing systems capable of detecting and correcting biases in media representation, which can transform media practices not only in Nepal but globally.
Vision for a More Inclusive Media Landscape
By making media more inclusive in Nepal, we’re taking essential steps toward fostering a more informed, empathetic, and united global community. This initiative not only aims to empower local media professionals with the tools to create more accurate and equitable media narratives but also sets a precedent for similar transformations in media practices worldwide.
Our Approach
Grassroot AI
Central to the concept of Grassroots AI is the focus on the human element in AI development, ensuring that solutions are technologically advanced, and culturally and contextually appropriate for the communities they are designed to serve. This approach involves community members in the development process, leading to AI solutions that are finely tuned to the specific needs and nuances of the local context. This inclusion fosters a sense of ownership and acceptance within the community, which is crucial for sustainable and impactful technology implementations.
In This Project:
We created a Comprehensive Media Analysis Tool to Improve Media Representation of Women, Youth, and Marginalized Groups
Step 1.
Engaging the Communityof Nepal to enable upskilling in the region andempower people through AI through Omdena Academy
Omdena Nepal Collaborators
In this initial phase, we introduced the community of beginners to theoretical introduction into the AI techniques that needed to be used to produce the final solution Natural Language Processing and Web Scraping, with a special focus on how these technologies can be used to analyze and understand media representation. Specifically, community members learned to extract large volumes of data from digital media sources, enabling the examination of how different issues and groups are portrayed in the Nepali media landscape.
Step 2:
Gaining real-life experience through open-source Projects in Omdena Local Chapters
Collaborators meeting in Nepal Local Chapter
With the theoretical foundation in place, the opening of local chapters allowed the newly trained members to apply their skills in real-world contexts. They undertook projects such as Web Scraping to gather media data and used Natural Language Processing tools to analyze content. This not only provided hands-on experience but also started generating insights into the biases and trends in media representation within Nepal.
You can hear first hand about this experience from one of our project managers here.
Step 3:
AI Innovation Challenge
During the AI Innovation Challenge, the focus shifted to deeper data analysis and model building. The team used the collected media data to build AI models that could more accurately identify and categorize media bias and representation trends. This step was crucial for developing a more nuanced understanding of media dynamics in Nepal, leveraging AI to pinpoint patterns that human analysts might overlook.
Step 4:
Top Talent Project
In the final phase, our top developers used the insights and models developed earlier to create a sophisticated, comprehensive analysis tool. This product was designed to continuously monitor and analyze media representation, providing ongoing insights into how different social issues and groups are portrayed. This tool aims to offer stakeholders, like journalists and policymakers, valuable data to foster more balanced and fair media coverage in Nepal.
Each step of the project not only advanced the technical skills of the community members but also directly contributed to a deeper understanding of media representation in Nepal, showcasing Omdena’s unique ability to merge community empowerment with targeted, impactful technological solutions.
The Outcome
This solution has transformed the landscape of media analysis, offering a seamless and impactful user experience that sets new benchmarks in data handling and processing. At the core of our system is a sophisticated yet user-friendly platform that ensures all operations, from data collection to analysis, are efficient and effective.
Key Features That Make a Difference
Gives a Rapid Overview of Specific Group’s Representation in Media: Our tool digs deep into online news, social media, and radio broadcasts, offering a panoramic view of the media landscape in the Nepali language.
AI-Powered Bias Detection: With the latest in AI technology, our system automatically pinpoints biases, ensuring that every voice has a chance to be heard fairly.
Efficiency at Scale: Our platform handles the heavy lifting of data processing, allowing for swift, accurate analysis without the need for manual oversight for nation-wide, in-time analysis.
Future-Ready Architecture: Built with growth in mind, our system is designed to evolve, accommodating new features and updates smoothly without downtime.
Inclusive, Accessible Design: Our user interface is a model of clarity and efficiency. It displays updated, actionable insights through intuitive graphs and maps that invite interaction and make complex data easy to understand. This real-time information is not just a collection of numbers but a vivid picture of media trends and biases, brought to life through sophisticated visualization tools.
Transparent AI: We demystify the AI process, ensuring that every user can trust and understand the how and why behind the data-driven insights. When the AI flags an article as a bias, the AI explains to the user the reason behind the decision, so the user can review and consent to what the AI has flagged.
Spreading the impact even further:
Our solution was a key highlight at the AI Conference on March 3, 2024, in Nepal, a major event orchestrated by prominent organizations including the Ministry of Education, Science and Technology, The Asia Foundation, the Robotics Association of Nepal, UKAid, the Nepal Academy of Science and Technology, and AECON.
Additionally, our team is actively collaborating with the Government of Nepal to craft a forward-thinking AI Policy, reinforcing our dedication to boosting AI literacy and infrastructure across the nation.
Other Applications of such Technology
Applied Methodology 1:
AI Powered Bias Detection
Can also be utilized in:
Human Resources and Recruitment: In HR, AI-powered bias detection can improve fairness in hiring practices by analyzing job descriptions, recruitment communications, and resume screening processes to identify and eliminate gender, racial, or age biases.
Financial Services: Financial institutions can use this technology to identify and prevent bias in lending, insurance underwriting, and customer service. Ensuring that algorithms do not unfairly discriminate against certain groups helps promote equality in financial decision-making.
Healthcare: In healthcare, bias detection can be used to analyze patient treatment plans, diagnostic tools, and patient interactions to ensure that all patients receive fair and equal treatment regardless of their demographic background.
Law Enforcement and Legal Services: AI can help identify potential biases in policing strategies, sentencing, and legal decision-making processes. This application can be crucial for ensuring fairness and equality under the law.
Education: Schools and universities can utilize AI to monitor and address biases in admission processes, grading systems, and educational content. This helps promote a more inclusive educational environment.
Advertising and Marketing: Companies can use AI to analyze their marketing campaigns and customer engagement strategies to ensure they are not unintentionally targeting or excluding specific demographic groups.
Public Policy and Government: AI can assist in analyzing policies and public service strategies to ensure they do not inadvertently favor or disadvantage specific groups of people.
Real Estate: AI technology can help real estate companies analyze their advertising and selling practices to ensure they are not discriminatory and comply with fair housing laws.
Technology and Software Development: Companies developing AI systems can use bias detection models to ensure their products perform equally well across different demographics, thus avoiding the propagation of existing biases.
Applied Methodology 2:
A Comprehensive AI-Powered Approach to Diversity and Trends Monitoring
Can also be utilized in:
Crisis Monitoring and Management: Organizations can use AI to track media coverage and public sentiment during crises or emergencies to manage communications more effectively and respond to public concerns in real time.
Brand Monitoring and Strategy: Companies can deploy AI to analyze how their brand is being discussed or portrayed in the media. This can inform brand strategy adjustments, help in managing brand reputation, and identify new marketing opportunities or potential threats.
Consumer Behavior Analysis: By monitoring how different demographic groups are portrayed or discussed in media, businesses can better understand consumer perceptions and behavior, which can inform product development, marketing strategies, and customer service approaches.
Political Campaign Management: Political campaigns can use AI to monitor the media representation of candidates, political parties, and key issues. This information can guide campaign strategies, debate preparations, and targeted advertisements.
Public Health Surveillance: Public health organizations can utilize AI to monitor media for discussions related to health behaviors, disease outbreaks, or public reactions to health policies. This can assist in public health messaging and intervention strategies.
Regulatory Compliance and Monitoring: In industries with strict regulatory requirements, such as pharmaceuticals or finance, AI can help monitor compliance by tracking how certain topics are reported or discussed in the media.
Intellectual Property and Copyright Enforcement: Companies can use AI to scan media for unauthorized uses of copyrighted or trademarked content, helping to protect intellectual property rights.
Detecting Fake News and Misinformation: By analyzing patterns and discrepancies in how news is reported across different sources, AI can help identify potential fake news or misinformation, which is crucial for news agencies, fact-checkers, and regulatory bodies.
Cultural and Social Research: Researchers can use AI to study cultural trends and shifts by analyzing media content over time, helping to understand societal changes and influences.
Advertising and Media Planning: AI can analyze content popularity and audience engagement, helping advertisers and media planners to make informed decisions on where to place ads for maximum impact.
Diversity and Inclusion Metrics: Beyond gender, youth, and marginalized groups, AI can help organizations track and improve their media representation concerning other diversity aspects, like disability, sexual orientation, or socioeconomic status.
Public Relations: PR firms can use AI to keep a pulse on media coverage related to their clients. Immediate access to such data helps manage public image and respond swiftly to potential PR crises.
Time Frame
The whole project took just 6 months,knowing we needed to:
Gather participants in a new location, organize and conduct courses also locally.
Build Local Chapters, and conduct Open Source Projects that involve among others gathering and preparing the data with our new AI adepts.
Conduct testing different approaches for technical solutions and parameters.
Build a functional product Dashboard.
Further Possibilities
To summarize this initiative we wanted to draw the vision on potential steps that could be taken to further expand the impact and capabilities of an AI-driven media analysis tool:
Expanding Geographically: Collaborating with international NGOs, governments, and media organizations to adapt and deploy the tool across different countries, tailoring it to accommodate local languages and media landscapes.
Customizing for Specific Sectors: Adapting the tool for critical sectors such as healthcare, education, or politics to analyze media representations and biases, aiding organizations in these fields to refine public perception and improve policymaking through data-driven insights.
Integrating with Educational Programs: Partnering with universities and educational institutions to incorporate the tool into journalism and media studies curriculums, equipping future professionals with the skills to identify and address media bias.
Developing a Public Access Interface: Creating a public-facing version of the tool that allows journalists, researchers, and the general public to conduct their own analyses, promoting greater transparency and awareness.
