Saudi Arabia's ALLaM Model Joins IBM Watsonx as a Top Arabic Language Generator

The announcement was made at the IBM Think event underway in Boston. (SPA)
The announcement was made at the IBM Think event underway in Boston. (SPA)
TT

Saudi Arabia's ALLaM Model Joins IBM Watsonx as a Top Arabic Language Generator

The announcement was made at the IBM Think event underway in Boston. (SPA)
The announcement was made at the IBM Think event underway in Boston. (SPA)

The Saudi Data and Artificial Intelligence Authority (SDAIA) announced on Tuesday that its ALLaM model, which generates Arabic text, was included in IBM’s leading watsonx platform.

The announcement was made at the IBM Think event underway in Boston.

This selection is testament to ALLaM’s advanced technical capabilities.

During its experimental phase, the model underwent rigorous testing against international standards for generative AI to ensure its readiness to compete with other models on watsonx, a platform widely used by developers around the globe.

Currently available in a trial version, ALLaM’s inclusion in watsonx allows for further professional evaluation. The testing will be instrumental in accelerating the release of the model's full capabilities and establishing it as a highly competitive force in the field of Arabic language generation.

The inclusion also aligns with Saudi Arabia's, specifically with SDAIA's, broader mission to promote the Arabic language on regional and global scale. The efforts focus on preserving the integrity of the language while promoting its use by enriching Arabic content in various fields, including technical, cultural, literary, scientific, and other humanities-based domains.

Ultimately, this initiative aims to leverage AI technologies and digital applications to foster cultural diversity and benefit all humanity, regardless of language, nationality, or educational background.

These efforts contribute to the goals outlined in Saudi Vision 2030, driven by Prince Mohammed bin Salman bin Abdulaziz Al Saud, Crown Prince, Prime Minister of the Kingdom of Saudi Arabia, and Chairman of the Board of Directors of SDAIA, to make the Kingdom a global leader in advanced technologies, including those associated with AI.

ALLaM is the first Saudi-developed AI system designed to answer user questions on different knowledge domains in Arabic.

The groundbreaking model leverages cutting-edge AI technology. Trained on a massive Arabic language dataset, one of the world's largest, and supplemented by English content, ALLaM ensures comprehensive responses.

Users can submit inquiries in text or audio format, and ALLaM will answer in the chosen format, drawing from the most trusted sources in the Kingdom and the Arab world.

The ALLaM model is the product of the SDAIA-IBM partnership. This collaboration is a significant milestone on the road to advancing Arabic language applications within generative AI, said Regional Vice President of IBM Saudi Arabia Ayman Al-Rashed.

"This cooperation unlocks the potential of Arabic language models for both public and private sectors, aligning with the cultural needs of the region," he added.

Al-Rashed further highlighted the broader impact of this project, stressing: "Companies can leverage these models to develop innovative services."

This latest development strengthens Saudi Arabia's position as a leader in AI technology tailored to the specific needs of the regional market, he went on to say.

Artificial intelligence experts, technicians, innovators, company presidents, and policymakers formed part of the IBM Think event.



Paris Olympics Expected to Face 4 Billion Cyber Incidents

A general view of the Olympic rings on the Eiffel Tower a day before the opening ceremony of the Paris 2024 Olympics, in Paris, France June 25, 2024. (Reuters)
A general view of the Olympic rings on the Eiffel Tower a day before the opening ceremony of the Paris 2024 Olympics, in Paris, France June 25, 2024. (Reuters)
TT

Paris Olympics Expected to Face 4 Billion Cyber Incidents

A general view of the Olympic rings on the Eiffel Tower a day before the opening ceremony of the Paris 2024 Olympics, in Paris, France June 25, 2024. (Reuters)
A general view of the Olympic rings on the Eiffel Tower a day before the opening ceremony of the Paris 2024 Olympics, in Paris, France June 25, 2024. (Reuters)

As the Paris 2024 Olympic Games approach, cybersecurity officials are bracing for over 4 billion cyber incidents. They are setting up a new centralized cybersecurity center for the Games, supported by advanced intelligence teams and artificial intelligence (AI) models.

Eric Greffier, the technical director for Paris 2024 at Cisco France, told Asharq Al-Awsat that the Tokyo 2020 Games saw around 450 million cyber incidents. He added that the number of incidents expected for Paris is at least ten times higher, requiring a more efficient response.

Greffier explained that a single cybersecurity center allows for better coordination and a faster response to incidents.

This approach has proven effective in other areas, such as banking and the NFL, where his company also handles cybersecurity, he added.

The Extended Detection and Response (XDR) system is central to the company’s security strategy.

Greffier described it as a “comprehensive dashboard” that gathers data from various sources, links events, and automates threat responses.

It offers a complete view of cybersecurity and helps manage threats proactively, he affirmed.

The system covers all aspects of the Olympic Games’ digital security, from network and cloud protection to application security and end-user safety.

In cybersecurity, AI is vital for managing large amounts of data and spotting potential threats. Greffier noted that with 4 billion expected incidents, filtering out irrelevant data is crucial.

The Olympic cybersecurity center uses AI and machine learning to automate threat responses, letting analysts focus on real issues, he explained.

One example is a network analytics tool that monitors traffic to find unusual patterns.

Greffier said that by creating models of normal behavior, the system can detect anomalies that might indicate a potential attack. While this might generate false alarms, it helps ensure that unusual activity is flagged for further review.