In a recent thought-provoking statement, Sam Altman, CEO of OpenAI (Inventor of ChatGPT), voiced his concerns about the immense power of AI and the potential risks it poses. Altman cautioned, “We’ve got to be careful here. I think people should be happy that we are a little bit scared of this. I’m particularly worried that these models could be used for large-scale disinformation… [they] could be used for offensive cyber-attacks.”

AI presents real risks, but it’s essential to remember that humans are the ones behind its creation. To effectively address and overcome these risks, we must prioritize the human element. Instead of giving in to fear, we should take proactive measures to revolutionize the field of AI and shape its trajectory for a brighter future. 

Grassroot AI development, gaining traction, is all about bottom-up collaboration, inclusivity, and varied perspectives. It utilizes collective intelligence from diverse communities to create AI systems that are fair, transparent, and make a positive societal impact. It allows those with direct knowledge of a problem to contribute, thus avoiding biases.

On the other hand, top-down AI development is dependent on centralized decisions and expertise. In such projects, a handful of experts might design and implement an AI system, possibly disregarding broader community input or their specific needs.

AI regulation, like the EU’s AI Act, should be mindful not to sideline bottom-up AI development. Legislations focusing solely on top-down rules risk overlooking the value of community collaboration and inclusivity in AI development. This might limit diverse input essential for societal needs and bias prevention in AI. Burdensome regulations might even stifle smaller, community-driven AI initiatives, restricting their valuable contributions to AI development.

Current Issues in AI Development

Although issues in the development of AI solutions can’t be limited to just three topic areas, we have identified what we feel to be the most relevant at present: 

• Bias and Discrimination: There is a growing concern about bias and discrimination in AI systems, as they can reflect and perpetuate societal biases and inequalities. Examples of bias can include biased training data, algorithmic biases, and unfair decision-making processes.
  
• Privacy and Security: The increasing use of AI involves the collection and analysis of massive amounts of data, raising concerns about privacy and security. Also, ChatGPT raises privacy concerns due to the risk of data exposure and misuse. The potential misuse of personal data, surveillance, and breaches in AI systems pose significant risks to individuals’ privacy and confidentiality.
• Job Displacement and Workforce Transitions: The automation potential of AI and advanced robotics raises concerns about job displacement and the need for workforce transitions. Certain job roles and tasks can be automated, leading to potential unemployment and socio-economic disruptions if proper measures are not taken to retrain and upskill the workforce. The future of employment: How susceptible are jobs to computerization?

Grassroot AI development has the potential to address many of the current problems in AI while empowering future leaders to drive positive change in the field. We have identified a four step process to enable Grassroot AI collaboration.

Educate: Grassroot AI empowers community leaders

One essential element of bottom-up driven AI development lies in education and community engagement, which sets the foundation for a developmental AI ecosystem. 

Erum Afzal, Head of Omdena Academy and a Ph.D. scholar at Justus Liebig University Giessen exemplifies the empowerment of individuals through bottom-up AI development. Her contributions to various Omdena Challenges, from anomaly detection on Mars to preventing online violence against children, have propelled her career. Overcoming complexity and ambiguity, Erum developed creative solutions through collaboration, flexibility, and a growth mindset. Education forms the foundation of bottom-up AI development, and its significance is exemplified through various initiatives. 

For instance, at Omdena Ideathons are organized, uniting hundreds of individuals in the Omdena community passionate about AI, as demonstrated in our collaboration with Druk Holding & Investments (the commercial arm of the Royal Government of Bhutan). In this virtual ideation over 50 impact hubs submitted their ideas to employ technology to address critical challenges, such as climate change, healthcare, energy, and other social good objectives in Bhutan.

Open-source projects play a pivotal role in facilitating this educational aspect and prototyping ideas through a culture of collaboration and dissemination of knowledge.

By placing education at the forefront of bottom-up AI development, we not only foster continuous learning and growth but also empower individuals to contribute their skills and knowledge toward creating meaningful change in Bhutan.

Innovate: Grassroot AI empowers localization

Building on the educational foundation, organizations should strive to encourage developers to innovate by developing real-world AI models. Through carefully curated challenges and projects, participants take advantage of the opportunity to apply their skills to complex problems faced by organizations and industries.

Omdena Case Study 1: Innovative machine translation of local language in Bhutan

In the case of Bhutan, we developed the first-ever machine translator for Bhutan’s local language Dzongkha.

The machine translates from Dzongkha to English, making Bhutan’s cultural heritage more accessible globally. The project’s outcomes contribute to Bhutan’s digital culture, creative industries, and IT while enabling deeper engagement between tourists and locals, fostering cultural exchange and understanding.

Related article: How Bhutan is Leveraging AI for Social Good Through Grassroots AI

Omdena Case Study 2: Monitoring mangrove deforestation in Tanzania  

Another example of localized AI development is where Omdena is partnering with the UK Foreign Commonwealth and Development Office (FCDO)  to upskill local AI engineers and help build and deploy an AI solution to monitor mangrove conservation in Tanzania Rufiji delta.

Omdena’s localized AI development in Tanzania’s Rufiji delta is empowering local AI engineers to monitor mangrove conservation using AI. This not only promotes sustainability but also sets a precedent for future AI projects leveraging local expertise to tackle environmental challenges effectively.

Omdena Case Study 3: Promoting equality and inclusive media in Nepal

Omdena is leading a project in Nepal with the Asia Foundation to combat media bias and misrepresentation of marginalized groups in Nepal by developing an AI-assisted scoring model. The project is part of CSM STAND (regional programs funded by the U.S. Agency for International Development (USAID)). Through analyzing online content with natural language processing and machine learning, the project aims to measure the diversity and representation of women, youth, and marginalized communities in the media landscape. The ultimate goal is to raise awareness of bias, promote equality, and create a more inclusive media environment, empowering marginalized communities in Nepal

Deploy: Grassroot AI empowers a product mindset

“The product must solve a problem that is meaningful to the customer.” – Eric Ries, Author of The Lean Startup.

And who is the customer in the NGO world? The communities and the people being served. However, in the NGO sector, it often happens that solutions are being built without real user feedback. This approach can result in ineffective solutions that do not meet the actual needs and preferences of the intended beneficiaries.

Read more via: How to Build and Implement AI Products in the NGO Sector

In several projects with Catholic Relief Services, we adopted a product mindset from the outset. The team developed advanced ML algorithms that aligned skills with job opportunities in Gaza and the West Bank. We optimized these ML algorithms by analyzing contemporary participation trends and identified the most efficacious strategies for five Water Sanitation and Health initiatives. Our team also expanded these ML frameworks to cover three food security programs. Furthermore, we introduced a cutting-edge algorithm to discern individuals at the highest risk of water insecurity.

Learn more: Omdena Talent Helps Catholic Relief Services Create and Scale AI Algorithms That Address Poverty

Scale: Grassroot AI enables local hiring

Scaling is a vital aspect of any bottom-up AI development approach. To scale solutions, organizations should actively seek to hire talented individuals within their network who have gone through the educational and innovation phases. By fostering a community of skilled practitioners, organizations like Omdena extend their reach, empowering more organizations, NGOs, and industries with AI capabilities. 

For example, in Tanzania over 200 local practitioners have been trained, which increases the chances that solutions can be maintained and improved by tapping into the local ecosystems instead of relying on an external partner to do the work.

The mission of Omdena’s Dar Es Salaam, Tanzania Chapter is to run open-source AI projects to solve challenges faced by the local community.

Learn more: https://www.omdena.com/local-chapters/dar-es-salaam-tanzania-chapter/

What´s next?

Explore AI with us on your next venture. Enhance your skills at Omdena Academy and tackle AI challenges to address industry and community issues. By supporting Omdena’s mission for scalable AI and local hiring, you contribute to lasting change. Let’s collaborate for a meaningful future.

Contact us: [email protected] 

Have a project idea? Book a 20 mins use case exploration call

You might also like