AI News

Follow the latest AI and automation news

22 May 20266 Dhuʻl-Hijjah 1447 AH

BBC Experiment Reveals Ease of Manipulating AI Chatbots

The BBC conducted an experiment demonstrating the ease of manipulating AI chatbots. A journalist convinced ChatGPT and Gemini that he was a world champion in sausage eating through a single article posted on his personal website. The experiment relied on simple content formulation rather than complex techniques. This experiment reveals how AI can be influenced by the information presented to it, emphasizing the need for verification before reliance. The journalist's straightforward writing style led to unexpected responses from the chatbots, reflecting weaknesses in the evaluation and verification mechanisms of these systems.

Business

22 May 20266 Dhuʻl-Hijjah 1447 AH

Even If You Hate AI, You Will Use Google AI Search

AI-generated answers are increasingly influencing the web, drawing users in with their convenience. This trend threatens original digital creativity, as many overlook content created by artists and thinkers. Reports indicate that reliance on AI responses may reduce engagement with unique works. These challenges necessitate that decision-makers consider ways to protect and sustain digital creativity amid AI usage. Clear strategies are needed to support artists and thinkers in this evolving landscape.

Agents

22 May 20266 Dhuʻl-Hijjah 1447 AH

OpenAI Tests ChatGPT Integration in Microsoft PowerPoint

OpenAI has begun testing the integration of its AI chatbot, ChatGPT, into Microsoft PowerPoint, enhancing user experience in creating presentations. This new feature allows users to generate slides and update existing presentations using natural language commands, streamlining the workflow significantly. This development marks a significant step in incorporating AI into productivity tools, potentially transforming how presentations are prepared in the future.

💡

California governor signs first US executive order to protect workers from AI job loss

What strategies can be implemented to protect workers from AI job displacement?

Models

OpenAI's ChatGPT Solves Long-standing Mathematical Problem

22 May 20266 Dhuʻl-Hijjah 1447 AH

Models

22 May 20266 Dhuʻl-Hijjah 1447 AH

OpenAI's ChatGPT Solves Long-standing Mathematical Problem

OpenAI announced that its ChatGPT successfully solved a mathematical problem that had puzzled researchers since 1946, utilizing one of its latest inference models. This achievement marks a significant advancement in AI capabilities, demonstrating the model's ability to generate original scientific ideas without direct human intervention, showcasing progress in artificial intelligence. This development is a crucial step towards enhancing the role of AI in scientific research, potentially opening new avenues for innovation in fields like mathematics and applied sciences.

Arab AI Scene

AI becomes the default for Saudi consumers as Deloitte’s 2026 Digital Consumer Trends Report reveals a decisive shift

22 May 20266 Dhuʻl-Hijjah 1447 AH

Arab AI Scene

22 May 20266 Dhuʻl-Hijjah 1447 AH

AI becomes the default for Saudi consumers as Deloitte’s 2026 Digital Consumer Trends Report reveals a decisive shift

Deloitte's Digital Consumer Trends 2026 Report highlights a pivotal moment in Saudi Arabia's digital evolution, with generative AI transitioning from novelty to necessity. A survey of 1,000 consumers reveals that 66% now use AI tools, a significant 17-point increase from last year. AI usage for work-related tasks has risen to 45%, with consumers increasingly relying on it for productivity. Common applications include information searching, idea generation, and translation, indicating a fundamental shift in how knowledge is accessed and problems are solved.

Models

Google I/O showed how the path for AI-driven science is shifting

22 May 20266 Dhuʻl-Hijjah 1447 AH

Models

22 May 20266 Dhuʻl-Hijjah 1447 AH

Google I/O showed how the path for AI-driven science is shifting

During the keynote at Google I/O, Demis Hassabis, CEO of DeepMind, stated that we are currently at the "feet of singularity." This remark came alongside a video showcasing how the WeatherNext program helped save lives during Hurricane Melissa in Jamaica last year. This contrast highlights the gap between the high ambitions of AI and the tangible outcomes of WeatherNext, which focuses on AI tools designed to solve specific scientific problems, while the latter aims for systems based on language models capable of conducting advanced research without human intervention.

Business

OpenAI opens Singapore AI lab as IMDA updates AI framework

22 May 20266 Dhuʻl-Hijjah 1447 AH

Business

22 May 20266 Dhuʻl-Hijjah 1447 AH

OpenAI opens Singapore AI lab as IMDA updates AI framework

OpenAI has announced the launch of its first Applied AI Lab outside the US in Singapore, backed by over S$300 million. This initiative, named OpenAI for Singapore, aims to create over 200 technical roles in the coming years. The lab will serve as a global hub for engineers working on AI deployment, aligning with Singapore's AI Mission priorities such as public service and finance. OpenAI plans to collaborate with local partners on education and workforce programs, including a chapter of the OpenAI Academy.

Models

HiDream AI Unveils Unified Multimodal Model with 200B+ Parameters, Raises New Funding

22 May 20266 Dhuʻl-Hijjah 1447 AH

Models

22 May 20266 Dhuʻl-Hijjah 1447 AH

HiDream AI Unveils Unified Multimodal Model with 200B+ Parameters, Raises New Funding

HiDream AI has launched the HiDream-O1-Image-Pro, a unified multimodal model based on the Unified Transformer architecture. Announced during the company's first open day on May 19, the model boasts over 200 billion parameters, enabling it to understand and generate images and videos. Unlike traditional models, HiDream-O1 integrates all capabilities from the initial training phase, allowing it to handle tasks such as text-to-image, image-to-text, and video processing within a unified framework. This new strategy marks a departure from common methods that separate understanding and generation tasks.

Models

ByteDance Releases Lance, a Lightweight Native Unified Multimodal AI Model

22 May 20266 Dhuʻl-Hijjah 1447 AH

Models

22 May 20266 Dhuʻl-Hijjah 1447 AH

ByteDance Releases Lance, a Lightweight Native Unified Multimodal AI Model

ByteDance has launched Lance, a lightweight unified multimodal AI model with 3 billion parameters. Unlike existing multimodal models that separate understanding and generation, Lance is designed as a single system for image and video understanding, generation, and cross-modal editing. This unified architecture allows it to operate efficiently across various media types, addressing the growing demand for compact AI solutions.