Open Source is key to developing AI.

Latin America, however, still has a lot of room to grow: ECLAC data indicate that the combined investment of all countries in the region in AI last year did not exceed 1.7% of U.S. investment or 5% of China. The difference is mainly explained by a still-developing digital infrastructure, which hinders the implementation of technology, and the gap in technical and data skills.

To overcome these challenges and accelerate the implementation of artificial intelligence, many organizations are betting on open source. With solutions offering everything from resilient and secure infrastructure to platforms for process automation, these technologies enable large-scale development, testing and execution of language models (LLMs) for business applications.

"The value that open source offers companies is to enable them to add domain-specific knowledge and skills to the models, customizing them to their use cases. In this way, they can start using GenAI in a more cost-effective, efficient and secure way"

explains Alejandro Raffaele, Enterprise Segment Leader for Latin America at Red Hat.Entry point

In practice, these technologies serve as an accessible and beneficial entry point for enterprises to adopt and implement artificial intelligence in their processes and operations more quickly, adapting AI functionalities for different use cases in various industries.

Red Hat, a global leader in open source enterprise solutions, has helped drive this positive revolution in the market by offering a broad portfolio of integrated solutions. It all starts with the recently released Red Hat Enterprise Linux AI (RHEL AI), a foundational model platform that enables users to develop, test, and deploy GenAI models in a more aligned way.

As a foundation for building a coherent and effective enterprise AI strategy, the open infrastructure based on RHEL AI provides customers with a platform for large-scale language models from the IBM Granite family that has recently become open source. This gives organizations the flexibility to align these models with their application requirements.

"GenAI is a great catalyst for innovation. However, current LLMs are often not very open, cannot be enhanced without specialized professionals, and demand high infrastructure costs. Red Hat Enterprise Linux AI changes this landscape by transforming the way enterprises use AI and by providing a quick and easy entry point to these technologies,"

explains Thiago Araki, senior director of technology for Latin America at Red Hat.

Red Hat Enterprise Linux builds on this open approach to AI innovation by incorporating an enterprise version of InstructLab, an open source project to advance large-scale language models used in generative artificial intelligence applications. Created by IBM Research and Red Hat, the InstructLab project is an open, cost-effective solution for improving LLM alignment that opens its doors to anyone who wants to contribute, even if they have little machine learning experience.

"With Red Hat Enterprise Linux AI and its open source approach, organizations can drive GenIA innovation internally with confidence and transparency, while reducing costs and eliminating barriers to entry"

said Adrian Cambareri, head of the RHEL business unit for the Americas at Red Hat.Gap solution

According to the executive, the lack of data science skills and the need for significant financial investments, including the acquisition of AI infrastructure or the consumption of AI services, are some of the barriers that open source technologies, such as RHEL AI, help to break down. The solution also contributes to other challenges, such as the complex process of matching AI models to specific needs, the difficulty of integrating AI into enterprise applications, and managing both these apps and the lifecycle of the models.

"To truly lower the barriers to entry for innovation with AI, companies must be able to expand the list of who can work on AI initiatives while keeping those costs under control. With InstructLab's alignment tools, Granite and RHEL AI models, Red Hat aims to apply the benefits of truly open source projects (accessible to all and replicable, fully transparent and open for contributions) to GenAI in an effort to remove these barriers," he concludes.