AI News

Follow the latest AI and automation news

21 Sept 202610 Rabiʻ II 1448 AH

US and China Discuss Alerting Each Other to AI National Security Threats

Officials discussed establishing a mechanism for reporting AI incidents that could threaten national security. This initiative aims to enhance cooperation between the two countries in the field of artificial intelligence, necessitating greater coordination to address security challenges. Such a measure is a significant step toward improving rapid incident response, reflecting both countries' commitment to strengthening cybersecurity amid the growing use of AI technologies.

Business

21 Sept 202610 Rabiʻ II 1448 AH

Trustworthy AI practices make organisations 15 times more likely to achieve strong ROI

A new report from SAS in collaboration with IDC reveals that organizations adopting trustworthy AI practices are significantly more likely to report strong returns on investment. These organizations were found to be 15 times more likely to report double the ROI from their AI projects. The report, titled "Data and AI Impact Report: The New Economics of Trust," highlights that organizations with robust governance and data quality practices consistently outperform their peers. It also emphasizes the emerging challenges as AI systems become more autonomous, necessitating stronger oversight and accountability.

Governance

21 Sept 202610 Rabiʻ II 1448 AH

UN says AI safeguards can’t wait for certainty

A United Nations scientific panel has warned that governments must rein in increasingly capable AI agents before their risks are fully understood. This warning comes in the UN's first major assessment of OpenAI's hack of Hugging Face earlier this year. The report arrives as leaders gather in New York for the UN General Assembly, while the US and China engage in talks about AI. UN Secretary-General António Guterres has urged governments to cooperate in addressing the threats posed by AI, stating that "the world cannot afford a race to the bottom on AI safety."

💡

5 Companies Using NVIDIA AI for Clean Energy

What challenges hinder clean energy adoption in the Arab world?

Governance

Building standards for the next phase of AI

21 Sept 202610 Rabiʻ II 1448 AH

Governance

21 Sept 202610 Rabiʻ II 1448 AH

Building standards for the next phase of AI

OpenAI has announced a plan to develop unified global standards for artificial intelligence, emphasizing the need for coordinated evaluation and reporting. The company aims to enhance AI safety by establishing standardized assessment mechanisms, which would promote transparency and trust in this technology. This initiative could lead to improved international collaboration in AI, contributing to the reduction of risks associated with its diverse applications.

Arab AI Scene

Saudi Arabia advances data and AI leadership, promotes responsible use for humanity

21 Sept 202610 Rabiʻ II 1448 AH

Arab AI Scene

21 Sept 202610 Rabiʻ II 1448 AH

Saudi Arabia advances data and AI leadership, promotes responsible use for humanity

The Kingdom of Saudi Arabia continues to strengthen its leading position on the global digital map, with data and artificial intelligence (AI) technologies serving as key pillars in achieving the objectives of Saudi Vision 2030. The Kingdom is adopting an integrated approach that balances advanced technological innovation with the development of national human capabilities while establishing comprehensive ethical governance frameworks for the responsible use of these technologies. With 2026 declared the Year of Artificial Intelligence, national efforts led by the Saudi Data and AI Authority (SDAIA) are highlighting the Kingdom's advances in developing and governing advanced technologies. This approach reflects the Kingdom's determination to employ smart technologies as a catalyst for inclusive growth and a driver of quality of life.

Business

Zebra Technologies researchers use AI and physics to replicate warehouse imperfection

21 Sept 202610 Rabiʻ II 1448 AH

Business

21 Sept 202610 Rabiʻ II 1448 AH

Zebra Technologies researchers use AI and physics to replicate warehouse imperfection

Zebra Technologies has announced the launch of the STRIPES dataset, aimed at enhancing the accuracy of computer vision models in storage environments. STRIPES relies on realistic simulations of damage and chaos in warehouses, improving models' object recognition capabilities. The dataset includes 30,000 images generated using AI, featuring 3,000 different box types, allowing for realistic wear and tear simulations. Gravity simulation techniques were also employed to arrange boxes in a manner that reflects real-world clutter.

Agents

AgiBot Ships AGILE 2.0 Sense–Control Model for Open-World Locomotion

21 Sept 202610 Rabiʻ II 1448 AH

Agents

21 Sept 202610 Rabiʻ II 1448 AH

AgiBot Ships AGILE 2.0 Sense–Control Model for Open-World Locomotion

AgiBot has announced the release of AGILE 2.0, an advanced model that integrates visual perception and terrain understanding. This update aims to enhance robot interaction in open environments, enabling simultaneous navigation and manipulation without delays. AGILE 2.0 features the removal of time splits between movement stages, allowing cameras to continuously drive planning and control, thus reducing jitter caused by delayed vision. The new model's capabilities were showcased through exciting demonstrations, including fire-hoop jumps and robots balancing on large balls.

Agents

Radiant D35 RISC-V Dataflow Chip Partners With PrimeBot for Embodied Edge

21 Sept 202610 Rabiʻ II 1448 AH

Agents

21 Sept 202610 Rabiʻ II 1448 AH

Radiant D35 RISC-V Dataflow Chip Partners With PrimeBot for Embodied Edge

At PrimeBot's product launch in Shanghai on September 20, Ricore was announced as a compute ecosystem partner for the consumer robot maker. This collaboration aims to integrate dataflow-architecture AI chips into humanoid robot inference. The Radiant D35 chip features a self-designed RISC-V core paired with a dataflow compute fabric, enhancing motion control and multimodal reasoning efficiency. This design promises improved energy efficiency and reduced latency.

Models

Unitree Opens Partial Weights for UnifoLM-WLA-1.0 After 64-Task Reveal

21 Sept 202610 Rabiʻ II 1448 AH

Models

21 Sept 202610 Rabiʻ II 1448 AH

Unitree Opens Partial Weights for UnifoLM-WLA-1.0 After 64-Task Reveal

Unitree Robotics has begun releasing partial weights for its UnifoLM-WLA-1.0 model, a 6-billion-parameter world-language-action foundation model for humanoid robots. The weights and datasets are now available on the project page, while other codes and datasets remain pending release. The model is designed to perform 64 different tasks, relying on real-robot data from Unitree and public sources. It includes household tasks like folding towels and washing clothes, requiring precise coordination between arms and legs.