AI News

Follow the latest AI and automation news

17 May 20261 Dhuʻl-Hijjah 1447 AH

New math benchmark reveals AI models confidently solve problems that have no solution

A consortium of 64 mathematicians has launched the new SOOHAK benchmark, which includes 439 handwritten tasks, of which 99 are unsolvable. Google's Gemini 3 Pro leads in solving research problems with a 30% success rate. However, no model has surpassed 50% in recognizing broken tasks. The SOOHAK benchmark reflects the gap between impressive results and the broader research skills lacking in AI systems. Improving performance requires more computing power, yet this does not aid models in recognizing that some problems have no solutions. The benchmark underscores the challenges AI faces in research domains.

Business

17 May 20261 Dhuʻl-Hijjah 1447 AH

AI agents face trust issues in ‘high-risk’ industrial sectors: experts

Industrial sectors in China are racing to integrate artificial intelligence to enhance efficiency, supported by government backing. However, experts warn that critical markets like healthcare and aviation may face high risks in transitioning to autonomous agents. AI is seen as a driver of the "Fourth Industrial Revolution," moving towards systems capable of independent execution within workflows, according to industry experts at the IDC conference. These transformations involve the use of AI across multiple domains, reflecting the Chinese government's commitment to fostering technological innovation.

Arab AI Scene

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

WTISD 2026: Intelligent connectivity powers Middle East’s AI economy, future roadmap

World Telecommunication and Information Society Day is celebrated on May 17 each year, emphasizing the importance of resilient networks, cloud platforms, and data centers. This day has gained significance in the age of artificial intelligence, where modern economies rely on connectivity as a fundamental element. In the Middle East, government investments in AI and digital governance are accelerating, enhancing the digital economy's capacity for growth. Under the theme "Digital Lifelines - Enhancing Resilience in a Connected World," the day highlights the increasing need for robust digital infrastructure.

💡

Latest open artifacts (#21): Open model bonanza! Gemma 4, DeepSeek V4, Kimi K2.6, MiMo 2.5, GLM-5.1 & others. On CAISI's V4 assessment.

What factors contribute to the performance gap between open and closed AI models?

Models

OpenAI co-founder Greg Brockman takes charge of product strategy

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Models

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

OpenAI co-founder Greg Brockman takes charge of product strategy

OpenAI has announced plans to integrate ChatGPT with its programming product Codex, marking a strategic shift in AI development. This integration aims to enhance performance and efficiency in natural language processing, allowing users to leverage the capabilities of both systems in a unified manner. This move is expected to strengthen OpenAI's competitive edge in the market, as executives seek to deliver more advanced solutions that meet the growing demands of users in programming and linguistic interaction.

Models

New benchmark shows Claude Mythos and GPT-5.5 can develop real browser exploits autonomously

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Models

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

New benchmark shows Claude Mythos and GPT-5.5 can develop real browser exploits autonomously

Researchers at Carnegie Mellon University have developed a new benchmark to assess AI agents' ability to exploit real vulnerabilities in Google's V8 engine. The benchmark reveals that Mythos significantly outperforms GPT-5.5, albeit at a cost twelve times higher. This benchmark is crucial for evaluating AI's capabilities in creating real browser exploits, providing precise data on performance and cost. It highlights the substantial gap between different models in this domain.

Models

New benchmark confirms AI video generators look stunning but still can't reason about the world

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Models

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

New benchmark confirms AI video generators look stunning but still can't reason about the world

A new benchmark named WorldReasonBench evaluates video generators based on logical and physical plausibility rather than image quality. ByteDance's Seedance 2.0 leads the pack, outperforming Veo 3.1 and Sora 2, with commercial models scoring about twice as high as open-source alternatives. Logical reasoning remains the most challenging category for all models, highlighting that the transition from pixel generators to actual world models has yet to occur. This gap underscores the disparity between visual quality and logical understanding.

Agents

For $1.3 million a month, OpenClaw founder Peter Steinberger runs 100 AI agents that code, review PRs, and find bugs

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Agents

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

For $1.3 million a month, OpenClaw founder Peter Steinberger runs 100 AI agents that code, review PRs, and find bugs

Peter Steinberger leads a three-person team managing approximately 100 Codex models for the open-source OpenClaw project, resulting in a monthly expenditure of $1.3 million on OpenAI API. Steinberger views these costs as an investment in research, aiming to understand how software evolves when coding expenses are not a barrier. This initiative allows him to explore new possibilities in software development. This experiment is a significant step in understanding AI's impact on software development, potentially transforming how developers operate in the future.

Business

Strengthening Singapore’s AI Future: A New National Partnership

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Business

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Strengthening Singapore’s AI Future: A New National Partnership

Google DeepMind has announced a partnership with Singapore to leverage advanced AI in health, education, and sustainability. This collaboration aims to tackle complex challenges, such as enhancing healthcare services and improving educational outcomes. The initiative aligns with Singapore's government efforts to foster technological innovation and address pressing societal issues. By integrating AI solutions, the partnership seeks to create impactful changes in various sectors.

Models

Finding the molecular switches behind new infectious diseases

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Models

16 May 202629 Dhuʻl-Qiʻdah 1447 AH

Finding the molecular switches behind new infectious diseases

Claire Bryant utilized the Co-Scientist platform to identify genetic triggers associated with emerging infectious diseases. This platform enables researchers to analyze genetic data more efficiently, contributing to a better understanding of how these diseases arise. This step is crucial for developing new strategies to combat future pandemics. Co-Scientist integrates artificial intelligence with genetic data, providing advanced analytical tools. Through this platform, scientists can identify genetic patterns that may lead to disease outbreaks, enhancing their ability to respond swiftly.