Trending AI news stories + paper
Today's Top AI news stories and papers
AK’s Substack is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
AI news stories
Apple's MGIE (MLLM-Guided Image Editing) allows users to edit photos using plain language descriptions without directly interacting with photo editing software. Developed in collaboration with the University of California, Santa Barbara, the model can perform various editing tasks such as cropping, resizing, flipping, and adding filters based on text prompts. MGIE interprets user commands, then generates edits by blending multimodal language models, enabling users to simply describe desired changes, like making a photo brighter or adding healthier toppings to a pizza, with the model executing the edits accordingly.
OpenAI has established a new Child Safety team aimed at preventing misuse or abuse of its AI tools by children, as evidenced by a recent job listing for a child safety enforcement specialist. This move comes amidst increasing concerns about the potential risks associated with minors using AI, prompting OpenAI to collaborate with Common Sense Media and implement measures such as requiring parental consent for underage users. With growing calls for guidelines and regulations on the use of AI in education, organizations like UNESCO are advocating for government intervention to ensure the safe and responsible integration of AI tools into educational settings.
Microsoft's AI efforts have shifted from Bing to Copilot, now the forefront of their AI initiatives, emphasizing productivity and creation over search. The Super Bowl ad for Copilot marks a strategic move towards promoting AI-driven creativity, reflecting Microsoft's pivot away from search competition with Google towards enhancing user experiences across its products through AI integration. Copilot's evolution into a standalone product signifies Microsoft's commitment to consolidating its AI under a unified brand, with plans for further expansion and integration into Windows to empower users and enhance their computing experiences.
Microsoft-backed OpenAI is developing a type of agent software designed to automate complex tasks by taking control of a user's device. These tasks include web-based activities like collecting public data about companies, creating itineraries, or booking flight tickets. These new assistants, often referred to as "agents," are expected to handle more intricate personal and work-related tasks with minimal human supervision. As of now, OpenAI has not responded to requests for comment on this matter.
The Biden administration has appointed Elizabeth Kelly, a top White House aide, as the director of the newly established AI Safety Institute at the National Institute for Standards and Technology, aiming to regulate and ensure the safety of artificial intelligence technologies. Kelly, who has been instrumental in shaping President Biden's tech and financial regulation agenda, will lead efforts to create "red team" testing standards by July to ensure the safety of AI systems for consumers and businesses. The institute's mandate reflects the administration's commitment to fostering trust and wider adoption of AI by establishing universal safety standards, a move seen as essential for unlocking the benefits of rapidly advancing technology.