• AI Watch
  • Posts
  • Apple's Launches AI Image-editing Tool That Works by Describing What You Want Edited

Apple's Launches AI Image-editing Tool That Works by Describing What You Want Edited

Google Rebrands AI to Gemini With New App; Apple Debuts AI for Easy Photo Edits; Archaeologists Decode Ancient Text

February 08, 2024

Your weekly source for the latest developments in Artificial Intelligence in Technology, Business, and Culture. 🤖

This week, OpenAI says it wants to transform ChatGPT into a comprehensive personal assistant, potentially impacting its partnership with Microsoft. Meta joins over 200 entities in a new U.S. association for safe AI development. Meanwhile, the FCC bans AI-generated robocalls targeting scams and misinformation. This is in connection to a month-long investigation of AI-generated robocalls mimicking President Biden’s voice that were used to discourage people from voting in New Hampshire’s primary. And Roblox introduces an AI-chat translator, enabling real-time text translations among players.

📰📱

Google Rebrands AI Assistant to Gemini With Advanced Capabilities and New Features

Google has rebranded its AI chatbot and assistant from Bard to Gemini, the name of the G’s suite of AI models that power the technology. Along with the rebranding, Google introduced a new Android app for Gemini and integration within the iOS Google app. The company aims to evolve Gemini into a comprehensive AI agent that can assist users in various tasks, including integration with Google Search. This ambition reflects a broader trend among tech giants to leverage AI for productivity enhancement.

Google also launched Gemini Ultra 1.0, a subscription for access to its most advanced AI model for $19.99 per month. The service, which includes a two-month free trial, is part of Google One and is currently available in English in over 150 countries, with plans to expand language support. Gemini's capabilities extend to Google Workspace and Google Cloud, enhancing business productivity. Google's vision for Gemini includes making AI agents more context-aware by integrating user data from Gmail, Docs, and Drive, ultimately aiming for AI to perform tasks like scheduling, booking travel, and generating content with a high level of autonomy and personalization.

🤖👉 For details on Google’s commitment to developing AI assistants capable of performing more everyday tasks, read the full article [HERE].

📰 📸

Apple Unveils AI Image-editing Tool That Works by Describing What You Want Edited

Apple has launched a new AI Model called MGIE to upend the photo editing industry. This model allows users to modify images by using simple text prompts, which eliminates the need for traditional photo editing software like Adobe’s Photoshop. MGIE can perform a range of edits, from basic adjustments like cropping and filtering to more complex alterations like changing object shapes or enhancing brightness. MGIE works by interpreting the user’s prompts and then visualizing the desired result. In one example, an MGIE user added vegetable toppings to a photo of a pepperoni pizza and increased the contrast to make it brighter.

This innovation represents a significant leap forward for editors, offering a more intuitive and efficient method for photo editing. The MGIE model is available on GitHub and Hugging Face Spaces. Although Apple has been less prominent in the generative AI space, this development opens up new possibilities for vision-and-language research and applications.

🤖👉 For more details on Apple’s AI offerings and CEO Tim Cook's vision for the future of Apple's technology, read the full article [HERE].

📰📜

Archaeology Team Uses AI to Decipher the Meaning of Ancient Scrolls

A team of student researchers has deciphered Greek writing inside a 2,000-year-old scroll from an ancient luxury Roman town buried during the 79 CE volcanic eruption of Mount Vesuvius. Using machine learning algorithms trained on high-resolution scans from archived work, the team revealed a previously unknown philosophical work discussing music, taste, and colors.

This significant breakthrough in deciphering text, mostly from the philosophy of Epicurus, was part of the Vesuvius Challenge contest. An innovative competition that leverages artificial intelligence to uncover texts from the Herculaneum scrolls. This method opened a new chapter in studying ancient texts, proving that even the most fragile artifacts can yield their secrets without being disturbed. And could potentially lead to the exploration of other ancient texts. With the Vesuvius Challenges’ plans to expand the competition in 2024 to decipher at least 85% of one scroll, the future of archaeological discovery and the study of ancient civilizations looks brighter and more accessible than ever before.

🤖👉 To see more of the Vesuvius Challenge contest objectives and get a closer look at AI within archaeology, read the full article [HERE].