•
January 11, 2024
•
6 mins
Hello Niuralogists!
Step into this week's edition as we navigate the dynamic realm of artificial intelligence to present you with the most recent breakthroughs. Our primary focus is to dissect the implications of these updates on various aspects of our lives, ranging from workplaces and businesses to policies and individual experiences. In this issue, we'll unveil compelling advancements, featuring highlights like Premier AI Showcases at CES 2024 and OpenAI's Vigorous Response to the Copyright Lawsuit Initiated by NYT.
For deeper insights, continue reading...
CES 2024 in Las Vegas is currently showcasing an array of groundbreaking AI products and features at the world's largest consumer tech show. Notable highlights from the first two days include Holoconnects' Holobox, offering a glimpse into the potential future of digital communication through AI. Baracuda's smart mirror utilizes AI and natural language processing to analyze expressions, adapting to users' moods with light therapy, meditations, and affirmations. Samsung's smart robot vacuum now incorporates AI for stain detection and surface differentiation. Rabbit introduces the Tamagotchi-style r1 AI companion device, streamlining digital tasks with AI agents. Samsung's Ballie, a spherical home assistant, takes on various household tasks, including a projector for videos and calls. Invoxia's Minitailz, an AI-powered pet wearable, monitors pet vitals and alerts owners to potential health issues. LG's Rosie, a charming home robot AI agent, aids with chores and seamlessly integrates with LG's smart home appliances.
OpenAI has strongly responded to The New York Times's recent copyright lawsuit, asserting its practices and questioning the motives behind the legal action in a published blog post. The key points outlined include that the NYT opted for OpenAI's content removal option in August, only to file a lawsuit months later. The blog post also contends that the Times deliberately selected prompts to induce regurgitation, rather than reflecting normal customer usage. Additionally, OpenAI emphasized in a submission to the UK’s Communications and House Select Committee that current copyright laws pose challenges for developing models like ChatGPT without utilizing existing works. The significance of this response lies in OpenAI's firm stance against the Times' lawsuit, highlighting the ongoing debate around AI content usage and the potential impact on public perception, regardless of the legal outcome.
McAfee has unveiled an innovative AI-powered deepfake audio detection technology, named Project Mockingbird, at CES 2024. This proprietary solution is designed to safeguard consumers against the escalating threat of cybercriminals utilizing fabricated AI-generated audio for scams, cyberbullying, and manipulating public figures' images. In response to the challenges posed by generative AI tools, which enable convincing scams and voice cloning, McAfee Labs developed an industry-leading AI model as part of Project Mockingbird. The technology employs contextual, behavioral, and categorical detection models, achieving an impressive 90 percent accuracy rate. With concerns about deepfake technology on the rise, especially in the context of elections, McAfee's Project Mockingbird aims to empower consumers to distinguish between authentic and manipulated content, thereby enhancing digital security.
MIT's Improbable AI Lab, a division within the Computer Science and Artificial Intelligence Laboratory (CSAIL), has introduced a new multimodal framework named "Compositional Foundation Models for Hierarchical Planning" (HiP). HiP utilizes three different foundation models, each trained on various data modalities of language, vision, and action, to help robots create and execute detailed plans for household, construction, and manufacturing tasks. Unlike previous multimodal models, HiP does not require access to paired vision, language, and action data, making the reasoning process more transparent. The system, developed by MIT researchers, could enable robots to perform complex tasks such as household chores and multi-step construction activities, enhancing their adaptability and decision-making abilities in various scenarios. The CSAIL team tested HiP on three manipulation tasks, outperforming comparable frameworks, and believes the system has potential applications in real-world tasks in robotics.
OpenAI has officially announced the upcoming launch of its GPT Store, set to debut next week. Initially planned for November but delayed due to internal leadership changes, the store will empower users to easily share, discover, and monetize their custom GPT creations. OpenAI is considering a revenue-sharing model based on community usage, providing creators the opportunity to profit from their unique GPT models. The store will feature a leaderboard highlighting top-performing GPTs and showcasing the most valuable creations in various categories. This move is expected to mark a significant development in the distribution of AI models, akin to an "App Store moment" for GPTs, offering creators a platform to showcase their innovations and fostering a new era of AI builders on OpenAI's platform.
📬 Receive our amazing posts straight to your inbox. Get the latest news, company insights, and Niural updates.
Researchers at North Carolina State University have introduced a novel soft robot design named "twisted ringbots," capable of simultaneous rolling, spinning, and orbital movement without human or computer control. These robots, made of ribbon-like liquid crystal elastomers, contract and roll on surfaces above 55 degrees Celsius. The ribbon's contraction induces rolling, while the exposed portion remains stationary, creating forward motion. Additionally, the robot spins along its central axis, resembling a record on a turntable, and moves in an orbital path around a central point. The twisted ringbots, showcasing physical intelligence, hold promise for mapping unknown environments and navigating confined spaces. The researchers can fine-tune their behavior by engineering the device's geometry, offering potential applications in various fields, including soft robotics and environmental mapping. The study was published in the Proceedings of the National Academy of Sciences.
MIT researchers, with the assistance of an artificial language network, have identified the types of sentences that activate the brain's key language processing centers. The study reveals that sentences with greater linguistic complexity, resulting from unusual grammar or unexpected meaning, generate stronger responses in language processing centers. Sentences that are extremely straightforward or nonsensical elicit minimal engagement. By using an encoding model, the researchers trained an artificial intelligence model to predict human brain responses to new sentences based on the responses of the artificial language network. The study provides insights into how linguistic complexity affects brain activity during language processing.
Midjourney has honed its proficiency in generating remarkably realistic or, in the case of Ethan's examples, eerily unsettling interior spaces through extensive training on real estate images. With a mere prompt, users can effortlessly visualize customized room designs in a matter of seconds. The process involves opening Midjourney, typing /imagine, and providing specific details about the desired room. For example, a prompt like "a photo of a modern high-end bathroom with luxury marble floors and granite countertops, a double vanity, a jacuzzi tub, toilet, and warm, soothing lighting" can swiftly bring the envisioned space to life. Users can then tweak and remix the generated output as needed, or explore more unconventional ideas, such as Ethan's suggestions, which include scenarios like a modern living room with a bottomless pit.
.png)