Meta's AI assistant has finally arrived on Mac, but it needs to catch up in several areas.

      The newly introduced desktop assistant can interpret what is displayed on the screen and accept voice commands throughout macOS.

      Meta has now launched its Meta AI assistant for Mac with a specific desktop application, offering users yet another AI chatbot to keep open alongside their other tabs. The app is capable of analyzing a shared window, responding to inquiries about what’s displayed, creating content, and enabling voice dictation on macOS.

      Meta AI can actually perceive what appears on the screen.

      The most notable feature of the Mac app is its capability to observe a shared window and use the visual context for understanding. Users can inquire of Meta AI regarding elements on screen, receive suggestions, or have it assist in generating content based on visible information. Additionally, it provides dictation functionality across various applications, enhancing its utility as a desktop assistant compared to merely opening Meta AI in a web tab.

      Meta is also targeting creators and small businesses with this app. Its AI can link with Instagram and Facebook accounts, integrate with Meta advertising campaigns, and connect with Google Workspace, allowing it to analyze metrics like post reach, likes, shares, and saves, suggest content ideas, and produce presentations, documents, and spreadsheets. It can also manage recurring tasks such as weekly performance summaries.

      However, Meta has some progress to make.

      Mac users already have a variety of AI assistants available, and Meta’s initial desktop venture seems more like an effort to catch up rather than leading the way. Gemini can already engage with shared screen content, while ChatGPT and Claude provide more advanced computer control features that can actively interact with the Mac. Meta could still find a valuable niche by directly integrating its assistant with Instagram, Facebook, and its advertising tools, particularly for creators and businesses.

      For now, though, the Meta AI Mac application appears to be a promising initial step rather than an essential assistant. Meta has indicated that additional features will be introduced, potentially increasing its usefulness over time. The main challenge remains convincing Mac users that they require yet another AI assistant on their desktop.

      Varun is a seasoned technology journalist and editor with more than eight years in consumer technology media. His work covers...

      Windows 11 is eliminating the Drag Tray feature.

      Finally, the inconvenient Drag Tray feature in Windows 11 is being removed.

      Microsoft is discontinuing one of the more unconventional file-sharing experiments in Windows 11 after users expressed that the feature was frequently more aggravating than beneficial. The feature, known as Drag Tray, was intended to provide a faster method for sharing files. Users could grab a file in File Explorer, drag it toward the top of the screen, and a floating panel would appear with shortcuts for sending it to applications. While straightforward in theory, many users were already dragging files to that part of the screen for completely different reasons.

      OpenAI aims to monitor AI misuse without requiring customers to give up their data.

      Their new Private Safety Processing system adopts a different approach compared to Anthropic, which mandates a 30-day retention period for some advanced models.

      OpenAI seeks to enhance its ability to detect abuse of its advanced AI models without asking privacy-conscious customers to forgo Zero Data Retention. In an announcement on August 19, OpenAI unveiled Private Safety Processing, a system aimed at identifying suspicious activities across multiple interactions while maintaining ZDR for applicable API users. OpenAI states that its current ZDR measures evaluate requests on an individual basis, making it harder to recognize patterns across several interactions. Private Safety Processing is designed to address this gap without granting OpenAI employees access to the underlying prompts or responses.

      Google is offering students a complimentary year of AI Pro and a new Gemini Student Hub.

      Gemini is receiving additional tools to assist students with studying, organizing, and researching.

      Google is currently trailing behind OpenAI and Anthropic in the leading frontier model competition. Anthropic’s Claude Opus 5 and OpenAI’s GPT-5.6 Sol both surpass Gemini 3.1 Pro on key independent benchmarks, while Google’s highly anticipated Gemini 3.5 Pro is still under testing after reportedly not meeting internal performance standards. While Google aims to close that gap, it is also aggressively directing Gemini toward a distinctly different audience. Eligible U.S. college students can now access Google AI Pro for free for a full year, along with a new Student Hub specifically designed for studying, academic work, and research.