Día: 4 de noviembre de 2024

To assess each country’s preparedness for implementing Artificial Intelligence ethically and responsibly, UNESCO developed a tool called the Readiness Assessment Methodology (RAM).

In Latin America, 14 countries are currently implementing it, with Chile standing out as a global model, being the first to complete the assessment.

If there’s one area where consensus exists regarding Artificial Intelligence (AI), it’s the need for appropriate regulation to maximize its benefits and mitigate risks.

In 2017, Canada launched the world’s first national AI strategy, and according to the AI Index Report 2023 from Stanford University’s Human-Centered Artificial Intelligence Institute (HAI), 62 AI strategies have since been introduced globally, with 14 more in development, and 127 countries now have at least one AI-related law.

However, it was not until the 41st General Conference of the United Nations Educational, Scientific and Cultural Organization (UNESCO) in November 2021 that the world gained its first global framework for the ethical development of AI.

The rapid adoption of this technology and its impact on society had become a priority for UNESCO, which, in a 2019 plenary session, approved the creation of a global regulatory framework. In early 2021, UNESCO conducted a multidisciplinary consultation with global experts to establish principles and guidelines for the responsible development and use of AI. These principles were enshrined in the Recommendation on the Ethics of AI, a framework approved by UNESCO’s 193 member countries at the 41st Conference. “It is the first global instrument accepted and adopted by this many countries,” says Natalia González, UNESCO’s AI Ethics Coordinator for Latin America and the Caribbean.

The organization developed this recommendation as a guide to help governments, companies, and organizations better prepare to address, mitigate, and resolve the undesirable effects of AI. It covers 11 policy areas, including ethical governance, data policy, education, health, and research. The framework promotes AI literacy, ethical research empowerment in education and e-learning, and interdisciplinary research through investments and sector collaborations that respect human rights.

To implement the Recommendation on the Ethics of AI, UNESCO developed a concrete tool, the Readiness Assessment Methodology (RAM), designed to assess—through questionnaires and other participatory tools—how prepared a country is for the ethical implementation of AI across five dimensions: legal/regulatory, social/cultural, economic, scientific/educational, and technological/infrastructural.

A Practical Methodology

This tool helps countries evaluate whether their AI laws and policies are appropriate and aligned with UNESCO’s Ethical Recommendation principles, ensuring the positive development of AI that respects fundamental human rights.

RAM implementation involves forming national teams, hiring local consultants, and organizing events and workshops with stakeholders from the public and private sectors, academia, and social organizations.

Public sector participants use the RAM questionnaire to evaluate AI in the five mentioned dimensions, with the results guiding necessary institutional and regulatory changes. This information is complemented by data from the private and civil sectors gathered in interdisciplinary workshops. Finally, the AI Readiness Assessment Report is published, presenting RAM results and policy recommendations to address governance gaps and ensure a responsible AI ecosystem aligned with UNESCO’s Recommendation.

Globally, about 50 countries are working with this methodology, including 14 in Latin America. Chile was the first country worldwide to complete the RAM implementation and publish its final report. In the coming months, reports from Uruguay, the Dominican Republic, and Cuba are expected, while other countries continue their processes. “UNESCO has been using Chile and Uruguay as global examples due to their progress and leadership in this area,” says the international organization’s expert.

The Chilean Experience

Chile’s RAM implementation process comprised four phases and was completed in under a year. It was conducted by the Chilean consulting firm Foresight, contracted by UNESCO, with direct collaboration from Chile’s Ministry of Science, Technology, Knowledge, and Innovation. The project aimed to update Chile’s National AI Policy (2021), which was presented last May and incorporated all RAM recommendations. “The RAM process in Chile coincided with the release of the 2023 Latin American AI Index and the update of the National AI Policy. This placed Chile in a favorable position, allowing it to gather valuable information on its AI status and sparking discussions in areas previously overlooked,” explains José Guridi, co-founder of Foresight, the consulting firm responsible for RAM in Chile.

In the first phase, an interministerial commission was formed to diagnose AI within the state using the RAM questionnaire. Workshops were held across six macro-regions, engaging 300 participants from various sectors to gather insights on AI’s opportunities and challenges. This phase concluded with a roadmap reviewed by ministerial counterparts and an evaluation report with recommendations. “UNESCO’s RAM recommendations were invaluable in focusing efforts during the update of Chile’s National AI Policy (PNIA), advancing cultural and social impacts, establishing ethical principles, and promoting responsibility and transparency in AI development and use,” says Chile’s Minister of Science, Technology, Knowledge, and Innovation, Aisén Etcheverry.

In addition to informing the PNIA, the findings were incorporated into a bill presented to the Chilean Congress in May 2024. The bill promotes human-centered AI, aiming to protect health, safety, fundamental rights, and consumers. It proposes a combination of self-regulation and risk-based regulation, categorizing AI systems by threat level and aligning with the ethical principles of UNESCO’s Recommendation on the Ethics of AI, the minister notes.

The consultant indicated that a highlight of Chile’s experience was the participatory nature of the RAM implementation, enriching the results by considering public sentiment about AI. “In Chile, we found an overall optimistic view of AI, seen as a powerful potential to be harnessed,” Guridi explains.

The final RAM report recommends that Chile prioritize data protection and cybersecurity laws, formulate municipal-level AI strategies, evaluate AI’s impact on culture and the environment, attract investment in technological infrastructure, and mitigate workforce impacts through retraining plans. It also proposes adaptive governance, recommending the creation of a specialized body to oversee AI policy implementation, ensure alignment with current legislation, and suggest regulatory improvements.

The report is now publicly available and will become part of UNESCO’s AI Ethics Observatory, a platform for sharing best practices globally. Meanwhile, UNESCO will work with countries that have completed the RAM process to adjust and update methodologies. “Future dynamics will involve periodic review and adaptation of strategies due to rapid technological advancement,” says UNESCO’s Natalia González.

In response to AI’s impact on productive, educational, and commercial ecosystems, and to ensure transparency and accountability in its use, Brazil developed its National Artificial Intelligence Strategy in 2021, which was recently updated.

A key role in its implementation is played by the Brazilian Artificial Intelligence Observatory, which is collecting AI data across Brazil with a special focus on industry, government, health, and education, to provide updated information to various stakeholders, including policymakers.

Natural Language Processing (NLP) is a branch of Artificial Intelligence (AI) that enables computers to understand and translate text and audio in multiple languages, facilitating global communication. However, there is a shortage of tools and specific data to train dialogue systems in Brazilian Portuguese, a language spoken by over 200 million people. Addressing this issue is the Center for Artificial Intelligence (C4AI/USP) at the University of São Paulo, where more than 100 researchers are engaged in various innovative AI projects, including NLP for Portuguese and Indigenous languages, of which Brazil has over 150.

C4AI/USP is one of the 11 Research and Applied Engineering Centers in Artificial Intelligence established by the São Paulo Research Foundation (FAPESP) and promoted by the Brazilian government to advance AI across strategic areas in the country. These centers’ data and research will support the newly created Brazilian Artificial Intelligence Observatory (OBIA).

OBIA serves as a public, open-access repository designed to compile and provide information on AI developments in Brazil and to connect with other international observatories. “Its main goal is to track AI development, impact, and adoption in Brazil, both quantitatively and qualitatively. We aim to be an information repository for decision-makers, industry, and other stakeholders,” says Tuca-Luiz Alexandre Reali Costa, OBIA’s manager.

Launched in 2022 by Brazil’s Ministry of Science, Technology, and Innovation (MCTI) and led by the Brazilian Network Information Center (NIC.br), OBIA was identified as a top priority in Brazil’s National AI Strategy, approved in 2021 and updated in 2024. From the beginning, this policy aimed to foster entrepreneurship and technological innovation in Brazil through nine pillars, one of which is AI Governance, with OBIA as a flagship initiative.

Since then, a multidisciplinary team has worked on its development, launching its website in March and preparing for its public launch in September 2024.

The core team includes professionals from NIC.br and the Regional Center for Studies on the Development of the Information Society (Cetic.br), with support from the Center for Management and Strategic Studies (CGEE), the State System for Data Analysis Foundation (SEADE), and C4AI/USP, among others. “The International Telecommunication Union (ITU), UNESCO, the International Research Center on AI (IRCAI), and the OECD have also been crucial in aligning the Observatory’s efforts with global best practices,” adds Reali Costa.

The 11 AI centers contributing to OBIA cover strategic areas such as industry, health, cities, agriculture, and cybersecurity. These were selected in two public calls from 2021 to 2023 by FAPESP, the Ministry of Science and Technology (MCTI-MC), and the Brazilian Internet Steering Committee (CGI.br). An exception is C4AI/USP, the first of its kind to join the observatory, created in collaboration with IBM.

Four Vital Areas

The creation of the Brazilian Artificial Intelligence Observatory (OBIA) was a complex process with challenges such as keeping pace with rapid advancements, integrating multiple perspectives and methodologies, and establishing better coordination among data producers feeding the observatory.

One strategic milestone for OBIA was defining indicators across four key dimensions to analyze AI in Brazil. “At Cetic.br, we have been managing indicators on digital technology adoption in Brazil for over 20 years, and for the observatory, we’ve organized these into four segments: government, health, education, and various economic sectors. We will also incorporate academic training, knowledge production, and patents,” says Alexandre Barbosa, head of Cetic.br, affiliated with NIC.br.

To monitor this work, a multisectoral governance committee is being established to address the needs of both current and future indicators. “OBIA will draw on indicators from various sources, including biennial surveys and automatically updated databases. Although frequency will vary, a deep and updated analysis is guaranteed to maintain the observatory’s relevance and accuracy,” explains Reali Costa.

The repository will include AI-related documents from Brazil and other countries, such as national AI strategies from Chile, Germany, Japan, and others, which will be classified using a taxonomy to facilitate data retrieval and cross-referencing, according to Barbosa.

Challenges

Brazil stands out as a regional leader in AI, according to the latest 2023 Latin American AI Index, with strengths in infrastructure, human capital, data availability, and governance, as well as its extensive digitization of public services. Brazil is also the only Latin American country among the top 20 nations in terms of academic publications in AI.

However, challenges remain, including further enhancing AI infrastructure, particularly in remote and rural areas; expanding the use of this technology across economic sectors; and increasing the annual output of PhDs in the field, which is currently four times lower than in the United States. “We expect rapid progress with support from the AI centers promoted by the Ministry of Science and Technology,” says the head of Cetic.br.

To address these challenges, the 11 centers have already attracted public and private investments totaling 240 million reais through 2030, as stated in Artificial Intelligence: Mapping AI Centers in Brazil: Initiatives, Actions, and Projects (Panorama Setorial da Internet, No. 1, April 2024).

OBIA is also tackling some of these challenges by providing crucial information for decision-making. The observatory is currently working on launching a data visualization portal, integrating more AI centers—both public and private—and disseminating initial findings that will provide an up-to-date and concrete overview of AI in Brazil across various areas.

“Our ultimate goal is for OBIA to become a benchmark, cooperative, and multidisciplinary initiative that provides reliable, comprehensive knowledge to inform society and guide policies, strategies, and actions to promote the responsible development and use of AI in Brazil,” says OBIA’s manager.

The “Ethical, Responsible, and Transparent Algorithms” initiative, led by GobLab, the public innovation lab at Adolfo Ibáñez University, promotes the integration of ethical principles into public procurement algorithms that utilize AI or data science.

This pioneering project in Latin America has developed replicable tools to enhance transparency and prevent discrimination in pubArtificial Intelligence (AI) is a branch of computer science that uses algorithms and data to make autonomous predictions and decisions. Algorithms direct the machine’s actions, while data allows it to learn and improve its performance. To date, several countries have adopted AI for managing public services to make processes more efficient, such as those related to public administration.

One area where this technology has been applied is in social services, including the selection of beneficiaries for financial aid, managing hospital waitlists, and school admissions. But who ensures that these systems are free from gender, racial, ethnic, or social biases? Who explains to a mother why her child wasn’t selected for her preferred school and what data the AI used? Or to a business why its project was rejected by the state? This raises another key question: how is personal data from applicants protected?

In response to these concerns, the GobLab public innovation lab at the Adolfo Ibáñez University School of Government launched the Ethical, Responsible, and Transparent Algorithms project, introduced in Chile in 2020. Coinciding with the late 2019 launch of fAIr LAC—an initiative to promote ethical AI in Latin America and the Caribbean—GobLab’s director and project lead, María Paz Hermosilla, coordinated discussions with the Inter-American Development Bank’s innovation lab, BID Lab, to secure funding and address these challenges. Public procurement became the starting point.

“We found it intriguing to focus on public procurement. We saw an opportunity to foster public-private collaboration, raise awareness, and build capacity in ethical AI, while developing concrete tools to ensure a positive social impact,” explains BID Lab lead specialist Carolina Carrasco.

Thus, after ChileCompra formally joined the project in 2020, additional collaborators came on board, including the Digital Government Division, the Council for Transparency, the Ministry of Science, and the startup accelerator Magical.

Four years since its inception, this pioneering initiative in Latin America has developed innovative, replicable tools for the entire region.

Citizen Protection

A milestone in the Ethical, Responsible, and Transparent Algorithms project—and a first in Latin America—is the approval by the Office of the Comptroller General of standardized guidelines for public procurement of data science and AI services by the state, developed by ChileCompra as part of the GobLab initiative.

For the first time, these guidelines establish rigorous requirements for transparency, fairness, and data protection to ensure that public services can procure, for instance, predictive models, benefit allocation algorithms, or user personalization systems without bias or opacity. “This not only improves the efficiency of government acquisitions but also strengthens public trust in government management and fosters equal opportunities for suppliers and contractors,” says Hermosilla.

David Escobar, analyst in the Studies and Business Intelligence Division of ChileCompra, explains that the entity is responsible for establishing purchasing policies and guidelines, issuing regulations, and ensuring the protection of citizen information and transparency. “In the case of Ethical Algorithms, we seek to ensure that AI-related purchases consider bias reviews to prevent negative impacts on citizens, guaranteeing that public resources are allocated fairly and without discrimination,” he says.

Another pioneering tool, developed in collaboration with the Council for Transparency, is the “General Guide for Algorithmic Transparency,” for which over 800 public organizations were surveyed on their algorithm usage, and the integration of algorithmic transparency into Chilean regulation was studied. “With support from BID Lab, a draft was developed, piloted with seven algorithms across four public organizations, and the final regulation was published in 2023, marking the first Latin American regulation on this matter,” states the director of UAI’s GobLab.

Meanwhile, the Digital Government Division collaborated on the creation of the «Ethical Framework for Data Science Projects,» which provides tools for public officials to address legal and ethical challenges in AI systems. “Data is a strategic asset for public administration, enabling faster service delivery and anticipating user needs with tools like AI. For example, the Social Security Institute processes benefit payments directly by integrating data, eliminating the need for individuals to apply. Concrete tools, like those from this project, help institutions meet minimum standards, ensuring transparency and bias mitigation in algorithms,” says Kareen Scharmm, coordinator of Policies and Studies at the Digital Government Division.

Investors and Audits

The Ministry of Science, Technology, Knowledge, and Innovation contributed to developing the experimental methodology fAIr Venture, which assesses ethical risks and social impacts of AI technology investments. “Our role as a public service is to ensure that algorithms are fair, do not perpetuate biases, avoid arbitrary discrimination, and, fundamentally, do not violate fundamental rights. This is also important for the private sector, as they are the primary providers of these technologies for the public system,” says Alondra Arellano, AI advisor at the Ministry.

The second phase of the project, initiated in July 2023 with funding from the National Agency for Research and Development, will extend until 2025. The GobLab team has already launched two tools: an “Algorithmic Transparency Report Card” and a “Statistical Bias and Fairness Measurement” that will allow for algorithmic audits of systems in use in the public sector. “Specific tools will be used to evaluate developing algorithms, enabling institutions to determine their transparency and detect biases that require adjustment. These tools will be applied across various pilots, ensuring algorithms meet minimum regulatory standards,” says Scharmm.

Although BID is no longer formally involved in this second phase, Carolina Carrasco notes that they will continue to support the project through fAIr LAC, as they are convinced that fostering this type of collaborative work can accelerate social impact, improve quality of life, and promote economic growth in the region. “The UAI Ethical Algorithms project is a remarkable example. They have already demonstrated a clear path with tools applicable in any geography. With fAIr LAC, we are now looking at how to facilitate the adoption of these tools in other countries across the region.

The Humboldt Cable is the first submarine fiber-optic route that will connect Chile with Australia, enabling faster, more stable, and cost-effective connectivity between South America, Oceania, and the Asia-Pacific region.

This joint initiative between Google and the Chilean government aims to boost the digital economy of Chile and all of Latin America, while linking them to key global markets.

Fiber-optic cables form the backbone of the Internet by transmitting millions of data at high speed, with greater bandwidth capacity and low latency, meaning minimal delay in data transmission. This ensures smooth performance for critical real-time applications, from online gaming to remote monitoring of industrial processes, surveillance systems, financial transactions, and telemedicine.

Today, improved technological infrastructure is foundational to the development of advanced technologies such as Artificial Intelligence (AI), cloud computing, and services like 5G and the Internet of Things (IoT), all of which are essential for the digital and economic transformation of countries due to their significant impact on industrial, digital, and scientific productivity.

In this context, the construction of the 12,000-kilometer Humboldt submarine cable route, promoted by Google and the Chilean government through the state-owned company Desarrollo País, is considered a true revolution in local connectivity. The project will provide direct connectivity between South America and the Asia-Pacific region, reducing latency (accelerating data transmission) and granting greater autonomy and resilience to Latin American telecommunications, solidifying Chile’s position as a digital hub and technology center.

This public-private initiative is an example of innovative and efficient governance, involving multiple stakeholders who have demonstrated a long-term vision and strategic ability by partnering with Google, which has enabled the realization of this critical infrastructure for the country.

“Chile’s governance in the project has been fundamental in showcasing its value and attracting a range of stakeholders. This joint effort between Google and Desarrollo País aligns the long-term strategic objectives of both entities. For Desarrollo País, this new digital infrastructure positions Chile as a technology hub, facilitating the development of data centers and new technological solutions. It also fulfills the country’s vision of being the first in the region to connect directly with Asia and Oceania. We chose Google as a strategic partner for its extensive experience in deploying submarine cables worldwide, which guarantees serious execution with a world-class company,” explains Patricio Rey, CEO of Desarrollo País, the state-owned company responsible for managing long-term infrastructure in Chile.

The ambition to establish a direct fiber-optic link between South America and Asia originated in 2016, during Michelle Bachelet’s second term, initially considering China. In 2018, under the second administration of Sebastián Piñera, Chile’s Undersecretariat of Telecommunications (Subtel), with support from the Development Bank of Latin America and the Caribbean (CAF), commissioned studies on technical, legal, financial, and economic aspects, defining the Valparaíso-Sydney route as optimal. In 2021, Desarrollo País assumed the project, launching an international call the following year to find strategic partners for the cable’s construction. Google joined in 2023, showing a strong interest in collaborating with the Chilean state, sealing this public-private partnership under Gabriel Boric’s administration.

“This cable consolidates Chile’s position as the center of digital activity in South America, opening opportunities for new industries, jobs, and improved working and living conditions for thousands of people. Previously, there was no connectivity of this kind from south to south; this advancement is also geopolitically significant and should make us proud,” said Chile’s President Gabriel Boric during the initiative’s launch in January 2024.

Holistic Infrastructure

Fiber-optic cables connect continents and markets, driving the relentless rise of digital technologies, which in 2021 saw a $34 billion increase in annual export value across six Latin American economies (0.8% of total GDP), according to Google’s 2022 Digital Sprinters report. This value could quadruple, reaching $140 billion by 2030.

The Humboldt Cable will facilitate unprecedented technology exchange across the South Pacific, opening new doors for connections between South America and Asia’s major tech centers in Hong Kong, Tokyo, and Singapore.

This connection will be part of Google’s integrated infrastructure in Chile, which already includes a data center in Quilicura (2015), the Curie submarine cable (2020)—connecting Chile, the United States, and Panama—and a cloud region launched in Santiago (2021), the second in Latin America after Brazil.

“Chile has pursued a significant investment attraction policy and built digital capacities, complemented by the high availability of renewable energy. That is why Google has invested heavily in this country, establishing its only data center in Latin America and the Southern Hemisphere in Quilicura. Chile’s economic and political stability, as well as its extensive network of trade agreements, allow this center to provide services to the entire region,” says Nicolás Schubert, Government Relations and Public Policy Manager at Google Chile.

Patricio Rey also highlights Google’s digital infrastructure expertise as a key aspect of this public-private initiative, “which increases investment efficiency for the state and generates significant economies of scale,” he says.

While Chile will undoubtedly benefit, the Humboldt Cable aims to support all of South America. “The connection from Valparaíso benefits the Southern Cone and the region in general, allowing other companies to use the open cables. This facilitates new connections between data centers in Chile and Asia, creating expansion opportunities for telecom and tech companies. Currently, Chile’s submarine communications are primarily with the United States; this new connection opens access to Australia and Asia,” explains Cristian Ramos, Director of Infrastructure Development at Google.

Tech Hub and AI

According to Schubert, all of this will enhance users’ internet experience and be essential for real-time applications. “These advancements will drive Chile’s digital development and highlight it as a regional benchmark,” says Schubert, specifically referring to the development of technologies like Artificial Intelligence (AI) and cloud computing.

This is no small matter. The 2023 Global Interconnection Index predicts that 85% of global companies will expand multicloud access across multiple regions by 2025, with data traffic between the Asia-Pacific and South America projected to grow at an average annual rate of 28% over the next 20 years.

Thus, the Chilean government’s strategy is to leverage the Humboldt Cable to position Valparaíso as an emerging hub in South America. “Google’s commitment to Chile reflects the confidence that international technology investors have in our country, which is crucial for establishing it as a digital hub,” affirms Patricio Rey.

For Chile, its renewable energy advantages could promote data center installations, pushing cloud capabilities and digital services that will benefit large corporations, SMEs, startups, and emerging sectors. “Developers and creators will experience an enhanced environment. We saw this with cloud service investments. New infrastructures not only deliver direct benefits but also foster an ecosystem. This was evident with the Curie cable, which brought significant economic benefits to the country as a whole—a multiplier effect that allowed many other businesses and industries to become more efficient,” states Cristian Ramos.

Greater access to cloud computing will strengthen technologies such as IoT monitoring, which is already impacting companies of national interest, like Codelco. Chile’s state-owned copper mining company increased its production by 4% after implementing integrated operations centers and IoT technologies.

Scientific research will also benefit from the cloud. “Accelerated data transfer will impact astronomical observation in northern Chile, which generates large volumes of data that need to be transmitted to universities and research centers. Previously, this data took traditional northern hemisphere routes across the Pacific. Now, with a direct route, we expect significant improvements in speed and cost, benefiting science as well,” says Schubert.

Economic impact is another measurable benefit. According to Analysys Mason, Google’s submarine cables in Latin America and the Caribbean, spanning five countries, will generate a cumulative GDP increase of $178 billion from 2017 to 2027, creating approximately 740,000 additional jobs by 2027, representing an annual GDP increase of 1.08%. “For example, the Curie cable increased Chile’s international outbound capacity (more data traffic to and from other countries) by 30%. Between 2020 and 2027, the Curie cable is expected to contribute $19.2 billion and create 67,000 jobs in the region,” says Google’s Director of Infrastructure Development.

High-Level Engineering

According to the 2024 Cable Map, since 2018, Google has invested in 29 submarine fiber-optic cable projects, five of which connect Latin America: Monet (Brazil and the United States), Tannat (Argentina, Uruguay, and Brazil), Curie (Chile, Panama, and the United States), Junior (Rio de Janeiro and Praia Grande), and Firmina (United States, Brazil, Uruguay, and Argentina). These are high-level engineering projects, requiring significant investment and involving simulations, surveys, manufacturing, and seabed installation.

Humboldt, meanwhile, is under development. “This year, we are focused on obtaining permits for seabed studies, essential to plotting the cable route. We’ve completed a preliminary study using underwater mapping and now need to confirm the route with a detailed survey. Concurrently, we will begin the cable’s construction—a long and complex process involving the complete fabrication of the cable and the installation of repeaters. Once the route is finalized and permits are obtained, we will proceed with seabed installation, ensuring a service life of about 25 years,” explains Ramos.

The Curie cable’s landing in Valparaíso took place in April 2019. “The cable reached the shore and was covered with sand and anchored on land to prevent it from being pulled back to sea. The cable was pre-constructed in a factory and then laid on the seabed using specialized ships,” explains Ramos, illustrating what to expect soon with the Humboldt cable.

The Humboldt cable is scheduled for installation in 2025 and will be operational by 2026. “This will also be enhanced by Google’s Pacific Connect initiative—which connects French Polynesia, Fiji, Guam, the Mariana Islands, Japan, Hawaii, and the United States—strengthening network resilience and linking Pacific islands with continents. The more connection nodes, the more resilient the networks, allowing traffic to find alternative routes in case of outages,” concludes Cristian Ramos, Google’s Director of Infrastructure Development.

The University of Los Andes, the Humboldt Institute, and the SINCHI Institute in Colombia, in collaboration with Microsoft, have developed an AI platform that processes satellite data, bioacoustic recordings, and camera trap footage.

The project aims to monitor deforestation and biodiversity in the Amazon.

This innovative and powerful tool will provide key information to protect a critical ecosystem essential for global climate stability.

With 7 million km² and spanning nine countries, the Amazon is the largest tropical forest in the world, home to 10% of the planet’s biodiversity and responsible for delivering humidity and rainfall across South America, which helps stabilize the global climate. However, due to activities like illegal mining, agricultural expansion, overfishing, hunting, and deforestation, this unique ecosystem is under serious threat.

According to the 2022 Amazon Alive Report (WWF), 18% of Amazonian forests have been lost completely, and an additional 17% are degraded, with numbers expected to worsen, impacting biodiversity, global climate, and people—especially the 47 million residents in the Amazon region who rely on it for their livelihoods.

To help address these threats, in 2022, the project Guacamaya: AI for the Amazon emerged, backed by Microsoft’s AI for Good initiative, focused on promoting the use of Artificial Intelligence to address social and environmental challenges. The platform leverages AI to provide science and government with accurate, timely data on phenomena negatively impacting this unique global ecosystem. “Guacamaya emerged from a collaborative exploration between Microsoft and the University of Los Andes in Colombia. Through its Research and Training Center in Artificial Intelligence, CinfonIA analyzed how AI could address environmental challenges, including Amazonian deforestation. A key tool for this analysis was training an AI model that prioritized data quality. We sought collaborators with validated, high-quality data, leading to partnerships with the Humboldt Institute and the Amazonian Institute of Scientific Research, SINCHI,” explains CinfonIA researcher Andrés Hernández Celis.

The Amazon was chosen as a focal point due to its critical importance for global biodiversity, especially in Colombia, which has the highest biodiversity per square kilometer. “The Amazon rainforest influences climate patterns and provides water to regions like Bogotá and the páramos. However, deforestation has reached a critical point, threatening to turn this natural ecosystem into desert, which underscores the urgency of leveraging AI to tackle this goal,” Hernández adds.

This led to the development of real-time AI-based systematic monitoring and data delivery—capable of generating alerts to take action before the situation becomes irreversible.

AI with Environmental Impact

Guacamaya draws on three data sources from the Colombian Amazon: bioacoustic recordings, camera trap images, and satellite data on vegetation cover. Bioacoustic data is provided by the Humboldt Institute—Colombia’s leading biodiversity science organization and home to the region’s largest biodiversity repository; camera trap images are contributed by the Faculty of Sciences at the University of Los Andes, while satellite data is sourced from the SINCHI Institute and the Colombian Institute of Hydrology, Meteorology, and Environmental Studies (IDEAM).

Andrés Rengifo, Microsoft’s Director of Corporate, External, and Legal Affairs for the Andean South region, explains that by integrating aerial, acoustic, and camera trap data, the project achieves a 360º view of the Amazon, resulting in a richer and more accurate understanding of the environment. “Generative AI offers valuable benefits for decoding data and optimizing processes, and Guacamaya demonstrated this by connecting all the scientific information, traditionally isolated in universities or research institutes, on this platform,” he explains.

The scientific institutions contributed over 100,000 files, enabling CinfonIA and Microsoft to train algorithms capable of using sounds and images to detect the presence or absence of specific species. This provides valuable insights into the state of biodiversity in this part of the Amazon.

“AI training and refinement is an ongoing process with continuous data input, but these algorithms already had a strong foundation from similar analyses, allowing their adaptation to the Amazon ecosystem. The success of this project lies in the synergy created among diverse entities. Each institution brings its expertise and resources, creating a multiplier effect that yields innovative, effective solutions for complex issues like climate change and biodiversity conservation,” says Rengifo.

Microsoft fuels Guacamaya with algorithms and computing capacity essential for running intensive models that facilitate scientific research and human resources. “We have robust computational power on our Azure platform, the foundation for running these systems and storing information, allowing us to gain new insights and develop approaches to better understand this data. We contribute models, tools, computational capacity, and a highly specialized team dedicated to addressing social and human challenges through technology,” explains the Microsoft executive.

Revolutionary Methodology

The introduction of AI to studying the Amazon ecosystem has completely changed how information is gathered from this globally significant biological reserve.

To develop one of the most important natural sound repositories in South America over 25 years—collecting 25,000 audio files from over 1,300 species—biologists at the Humboldt Institute used limited-capacity recorders placed and retrieved from the forest. These were then delivered to experts in vocalization recognition, who spent months listening to audios and identifying species.

But the arrival of Guacamaya has revolutionized this field, and after a year and a half of work, the first results are evident. Today, AI algorithms not only identify bird sounds with 80% accuracy, analyzing data five times faster, but they can also detect threats like illegal logging and machinery. “In the bioacoustics field, a model was developed in collaboration with the Microsoft Speech Lab, converting sounds into spectrograms, which are visual representations of sound,” explains CinfonIA researcher Andrés Hernández. These spectrograms, combined with natural text (such as species names), help identify bird or insect sounds or specific frequencies, he notes.

The processing of camera trap images is also benefitting from the initiative. “Previously, a scientist manually reviewed over 100,000 images, but now Guacamaya does it automatically, conducting monitoring 10 times faster and saving 90% of the time. The algorithms filter useful images and identify specific animals, facilitating efficient analysis and the detection of rare or invasive species,” Rengifo explains.

Currently, the pre-trained models of the Guacamaya project are available on PyTorch Wildlife, Microsoft’s open-source platform for biodiversity, encouraging collaboration among developers in the field. “PyTorch Wildlife provided us with a collaborative deep learning framework for conservation, including various neural networks to tackle specific project tasks. An example is animal segmentation and detection in camera trap images using Microsoft’s Megadetector,” explains researcher Andrés Hernández.

Another significant contribution of Guacamaya is its ability to interpret satellite images of the Amazon forest in real time. According to official data from Colombia’s Institute of Hydrology, Meteorology, and Environmental Studies (IDEAM), the Colombian Amazon lost more than 40,000 hectares between 2021 and 2022. “Guacamaya uses high-resolution images to monitor forest loss, speeding up forest cover reports and enabling near-real-time conservation decisions. This contrasts with traditional reports, which in Colombia had an 18-month delay,” clarifies Rengifo.

Open Source

CinfonIA researcher Andrés Hernández highlights that the biggest technical challenge in training the algorithms has been data management. “For example, it is crucial to have detailed camera trap images to accurately identify species. This is difficult due to variations in the images and the uneven distribution of species in the database. Some species are common and overrepresented, while rare species appear fewer than 100 times. This imbalance presents a considerable challenge, as it requires managing an unbalanced data distribution.”

This explains why the mega data from each pillar—camera traps, sounds, and naturalist coverage—has been trained through either proprietary or pre-defined models, with bioacoustic and camera trap algorithms being the most advanced. “For satellite images, we are refining temporal consistency, which will enable reliable results at different times and avoid instability at forest cover map edges, thereby improving the accuracy of reforestation measurements. The main goal of our model is to expand current research by integrating the project’s three pillars into a platform to analyze the relationship between biodiversity loss, animal migration patterns, and deforestation across different regions,” Hernández explains.

To achieve scalability, Guacamaya was designed as an open platform for scientific collaboration, and in this new phase, following its official launch at the end of 2023, it aims to consolidate data on the Azure platform and invite more international institutions to participate in Amazon monitoring. “This project was designed to allow other members of the scientific community and countries like Peru, Ecuador, Bolivia, Brazil, Paraguay, Venezuela, Suriname, Guyana, and French Guiana to join and contribute. Open collaboration is essential for addressing challenges like climate change and gaining a global, comprehensive perspective on what is happening in this ecosystem,” states Microsoft executive Andrés Rengifo.

Adding data from diverse sources is therefore vital to continue improving algorithms and generating more in-depth and accurate analyses of the Amazon. To this end, the source code of the models and the platform is available to any research center or scientist interested in contributing to the protection of this area. “At Microsoft, we believe technology is a crucial tool for addressing challenges like climate change. We have received support from scientists and the Colombian government, who have provided data and shown enthusiasm in seeing how the platform operates and the potential of technological collaboration in biodiversity conservation,” says the Microsoft executive.

The Guacamaya team continues to work on improving the algorithms and expects to have functional models in each category to present the initiative’s progress at COP16 in Cali, Colombia. “This initiative is not just the responsibility of a single country or group; it is a joint effort that involves the entire Amazon region and beyond. Our goal is to expand this tool to institutions and organizations in other countries working in the Amazon to join the Guacamaya project. The Amazon is not only a natural treasure but also the planet’s lung that regulates the global climate. Protecting this invaluable ecosystem is a shared responsibility that transcends borders and requires a collective effort from all Amazon countries and the international community,” concludes CinfonIA researcher.

As a unique and groundbreaking service, one of Brazil’s largest educational companies, SOMOS Educação, is offering AI tools for teachers and students through its Plurall platform.

This includes Plu, the intelligent assistant developed in collaboration with Amazon Web Services (AWS) to support teachers with lesson plans that optimize their time.

Plu will be implemented as a pilot and represents the first of several company projects leveraging AWS’s generative AI tools, aiming to reach over 7,000 schools in Brazil.

On average, teachers spend two hours a day preparing a class, a routine that could soon change for many educators in Brazil with the help of an intelligent AI assistant capable of generating a detailed plan for a 50-minute lesson with just a request. This assistant, Plurall AI—known as Plu—is a generative AI tool created by one of Brazil’s largest educational companies, SOMOS Educação, in collaboration with Amazon Web Services (AWS), designed to help teachers and students plan lessons more quickly, accurately, and effectively.

At Bett Brasil 2024, the country’s premier educational fair, SOMOS Educação chose to unveil the pilot for this virtual assistant, capable of delivering a complete lesson script to a teacher in just seconds. Not only does it provide a detailed lesson plan, but it also includes illustrations, suggested activities for students, and even customized questions for students who need exams tailored to lower difficulty levels.

SOMOS Educação aims to revolutionize primary education in Brazil by reaching over 5,000 schools with the intelligent assistant Plu by 2025. The main goal of this technology is to enable teachers to dedicate more of their work hours to personalized student engagement. According to the OECD’s 2018 Teaching and Learning International Survey (TALIS) and McKinsey’s 2020 report, 67% of teachers’ time is spent on activities outside the classroom, such as lesson preparation. McKinsey estimates that 20% to 40% of these activities could be optimized with technology, which could provide professionals with up to 12 extra days of productivity annually for each 5% optimization. “We believe this technology can be incredibly valuable in freeing them from time-consuming tasks. If we consider a conservative optimization goal of 10%, it could mean nearly 24 additional days per year for teachers to support students, enhance lessons, or focus on their own well-being. These metrics may change over time and could lead to other measurable impacts,” explains Rafael Augusto Teixeira, Senior Computing Manager at SOMOS Educação.

“Generative AI is one of the most transformative technologies of our generation. It addresses some of humanity’s toughest challenges, enhancing human performance and maximizing productivity. (…) In education, AI has a significant impact, offering several advantages that can transform how we learn and teach,” states Cleber Morais, AWS Director of Enterprise Sales for Latin America

Successful Outcomes

For nearly a century, SOMOS Educação, a leader in primary and secondary education in Brazil, has been providing comprehensive services for schools, including educational tools and online learning resources. Since 2014, its digital platform, Plurall, has served over 7 million students and 120,000 teachers across 7,000 schools, offering digital books, activities, assessments, and other online tools to this community.

Aware that AI is here to stay and holds great potential to transform the educational landscape, SOMOS Educação approached the AWS team in June 2023 with the goal of launching a GenAI-based solution to impact the education market. “After collaborative efforts between AWS and SOMOS Educação teams, the virtual assistant for teachers was selected as the project to invest in, given its potential for scalable adoption,” explains Morais.

Thus, within its digital platform Plurall, the education company integrated this generative AI-based chatbot to provide advanced, personalized solutions. “Plu uses our extensive content base to generate responses and meet user requests, including creating complete lesson plans, text summaries, activity lists, illustrative images, unique questions, comprehensive exams, interdisciplinary lessons, bilingual content, and adjustments to the complexity level of certain topics,” Teixeira elaborates.

The intelligent assistant Plu is available to both teachers and students. For students, it can analyze content, generate questions, request summaries, provide additional activities, and even create study plans, among many other applications. For teachers, through the Adaptive Teaching tool, it can recommend new content and skills for students to work on. By July 2024, 3,400 students from various schools had been testing the assistant.

“Not only have school owners and administrators highlighted the platform’s positive outcomes, but teachers and coordinators have also praised it. The reception has been so enthusiastic that we created a waitlist for the pilot. Teachers with decades of experience have noted that the tool provides new perspectives, enhances class preparation, and supports blackboard summaries,” Teixeira states.

AWS Generative AI

To create the intelligent assistant Plu, SOMOS Educação utilized Amazon Bedrock, a cloud service platform from Amazon Web Services (AWS) that enables developers and companies to build, train, and deploy AI models tailored to a variety of specific tasks. By using prompts—tools to guide and customize responses—the developers adapted the behavior of these pre-trained models to meet the unique needs of the educational context. This customization improved the chat applications used by teachers and students, ensuring more precise and relevant responses to school-related requests.

Among the AWS tools used in the development of Plu were CloudFront for front-end content caching and RDS for managing school users and data. Other critical services included Amazon EKS (Elastic Kubernetes Service), which facilitates cloud application management and ensures the assistant operates efficiently and without interruptions; Amazon S3 (Simple Storage Service), providing secure and scalable cloud storage for quick access to essential information; Amazon OpenSearch Service, enabling fast and accurate data searches and analysis, ensuring teachers quickly access needed information; Amazon SQS, supporting message exchanges between different system components for smooth chatbot communication; and finally, Amazon SNS, which allows for efficient notification distribution to teachers and students, enhancing overall service communication and coordination.

AWS’s executive for Latin America noted that the company collaborated with Accenture in designing the GenAI model architecture for the virtual assistant, as well as in the development, implementation, and training of SOMOS Educação’s teams. “AWS offers companies more than just a chatbot, tool, or LLM: we provide multiple capabilities, including generative AI-powered applications, tools for creating custom generative AI applications, and scalable infrastructure. All with security safeguards and controls, allowing companies to operate with confidence. In the near future, all applications will feature generative AI to make them more useful, personal, and engaging,” says Morais.

One of the strengths of SOMOS Educação’s intelligent assistant is the company’s extensive educational content database—one of the largest globally by volume—which serves as the assistant’s foundational knowledge. To manage this resource, SOMOS relied on guidance from AWS Brazil and employed the RAG (Retrieve, Augment, Generate) approach, a method that supports AI—especially in language and text generation models—to enhance the quality and relevance of responses generated by AI systems like this chatbot.

“We believe that SOMOS’s vast AI-driven knowledge base can rapidly drive changes in the learning process, not only in Brazil but globally. This technology can be easily adapted to other languages and content bases, thanks to the power of AI,” states Teixeira.

New Tools

Currently, the implementation of generative AI through this virtual assistant represents SOMOS Educação’s cutting-edge offering in education. In fact, the company has established a long-term roadmap to integrate generative AI across the entire platform, explains Bruno Brusco, the digital director responsible for the entire Plurall platform operation.

In the first phase, the focus will be on enhancing teacher productivity through AI usage. In the second, specific support will be provided to both teachers and students through the creation of adaptive learning pathways, which students can use for personalized assignments or in learning games and challenges assigned by tutors. In the third phase, SOMOS aims to leverage AI-generated data for two key purposes: predicting students’ academic performance and providing insights to assist administrators in making informed decisions. This will involve detailed information on improvement areas, identified educational trends, and strategic recommendations to enhance student performance and educational management efficiency.

“We have the best educational content in Brazil, and combined with our strong investment in technology, this positions us as leaders in large-scale AI initiatives for education,” states Brusco.