AI This Week

OpenAI has introduced ChatGPT Atlas, an artificial intelligence-powered web browser built around its popular chatbot. Reuters reports that this launch directly challenges Google Chrome's dominance in the browser market. Atlas offers AI features like content summarization and task automation, potentially accelerating the shift towards AI-driven search. The browser allows users to open a ChatGPT sidebar in any window and includes an "agent mode" for paid users, enabling ChatGPT to interact with websites on their behalf. Atlas is currently available globally on Apple's macOS, with versions for other platforms coming soon

A new investigation probes whether artificial intelligence chatbots respond differently based on the language used. By asking varied questions in multiple languages — English, French, Chinese, Hindi, Arabic, and Spanish — the study tested platforms like ChatGPT, Claude, and DeepSeek. While responses in different languages showed subtle nuances, AI models exhibited consistent center-left, secular values regardless of linguistic context. For example, AIs unanimously rejected cultural biases favoring sons over daughters or violence in relationships, emphasizing egalitarian views. However, platforms like DeepSeek demonstrated unexpected shifts in tone for politically sensitive topics when switching languages. This raises questions about training data biases and the cultural imprints of large language models.

The Wikimedia Foundation warns of a sharp decline in human visitors to Wikipedia, driven by the growing use of generative AI chatbots and summary-driven search engines. These technologies utilize Wikipedia’s vast content without directing users to the website, challenging its sustainability. The Foundation argues this trend not only reduces visitor engagement but also threatens content development and funding from donations. Wikipedia remains a key resource for AI training and search engines, yet this reliance siphons traffic away from the site. Illustrating the problem, human pageviews on Wikipedia dropped roughly 8% compared to the same period last year. Efforts are being intensified to ensure responsible reuse of the platform’s content.

Scouting America has introduced two new merit badges — Artificial Intelligence and Cybersecurity — aimed at equipping Scouts with essential digital skills. The AI badge covers topics like machine learning, ethical decision-making, and real-world applications, while the Cybersecurity badge focuses on encryption, virus protection, and responsible online behavior. These badges are part of Scouting America's 2025 requirements updates and are available for Scouts to begin earning today. The initiative reflects the organization's commitment to preparing young people for success in an increasingly digital world.

Uber is piloting a system that lets its drivers and couriers earn extra money by completing “microtasks” for AI training — things like recording voice samples, uploading images, or submitting documents in different languages. These tasks normally live in the realm of large-scale AI labeling services like Scale AI or Mechanical Turk, but Uber wants to fold them into its massive, decentralized network of drivers. The move could change how AI gets trained and who reaps the rewards. Uber argues it can combine work and data generation into a unified app experience, but it also raises questions about fairness, compensation, and worker classification.

Microsoft has introduced MAI-Image-1, a new generative AI-powered photo generator. Developed in-house, this text-to-image tool creates photorealistic images and various artistic styles within seconds. Unlike Microsoft's Copilot AI, which uses OpenAI's technology, MAI-Image-1 is entirely Microsoft's creation. The model aims to provide flexibility and visual diversity for creators, competing with offerings from OpenAI, Google, and Meta. Microsoft emphasized rigorous data selection and evaluation focused on real-world creative use cases. Interested users can test MAI-Image-1 on Microsoft's official website, though no mobile app is currently available.

California has made history as the first state to regulate AI companion chatbots. Governor Gavin Newsom signed SB 243, imposing strict rules to safeguard children and vulnerable users. Starting January 2026, chatbots must include age verification, transparency about AI-generated content, and restrictions on inappropriate behavior, such as posing as medical experts or sharing explicit material. Companies like OpenAI, Replika, and Character AI must implement self-harm intervention measures and report crisis-related data to public health authorities. SB 243 also raises fines for illegal deepfakes to $250,000 per violation. The law responds to growing concerns about chatbot misuse, including leaked cases of troubling interactions. A parallel bill, SB 53, mandates greater transparency in AI development

A new Gartner survey shows 77 percent of customer support leaders feel pressure from executives to deploy AI across their organizations. In response, 75 percent report their budgets for AI initiatives have already grown this year. Many plan to hire additional full-time employees (FTEs) to manage AI investments. According to Gartner’s analysis, the most valuable AI use cases in service and support include agent enablement, automated operations, self-service, and agentic AI. The report signals a growing expectation that AI will be central to customer support strategy and execution.

Amazon has introduced Quick Suite, an enterprise AI “teammate” that pulls data from emails, files, databases, and apps to answer questions, build custom agents, and carry out tasks. AWS describes it as “everything you want to do with ChatGPT at work, but can’t,” and backs it with data protections. Admins connect Google Drive, Office 365, Slack, Salesforce, Jira, Snowflake, Redshift, Databricks, Oracle, and more. Users tap Quick Research for reports using company context, Quick Sight for analysis and visuals, and automation tools like Quick Flows and Quick Automate for repetitive and complex processes.

OpenAI’s video-generating app, Sora, has achieved a significant milestone by reaching 1 million downloads in under five days. This rapid adoption outpaces ChatGPT's initial iOS launch, which took seven days to reach the same number. Notably, Sora's invite-only launch in the U.S. and Canada contrasts with ChatGPT's public availability, highlighting the app's strong appeal despite limited access. Early user engagement has been robust, with Sora climbing to the No. 1 spot on the U.S. App Store and maintaining high daily download rates. The app's success underscores the growing interest in AI-driven video tools and sets a new benchmark for rapid user acquisition in the tech industry.