Anthropic has revamped Claude's guiding document, now a comprehensive 57-page 'Constitution' that defines the AI's ethical framework and core principles. Unlike the previous version, which was a simple guideline list, the new constitution aims for Claude to grasp the reasoning behind its behavioral expectations. It addresses how Claude should manage conflicting values and navigate complex, high-stakes scenarios, emphasizing a deeper understanding of its role. This refined approach reflects Anthropic’s commitment to instilling responsible and transparent AI behavior at the heart of their model.
Apple is reportedly developing a small AI-powered wearable device, roughly the size of an AirTag, equipped with cameras and microphones designed to capture the user's surroundings. The device will feature a thin, flat, circular design made from aluminum and glass, including a standard and wide-angle lens, three microphones, a speaker, a physical button, and wireless charging capability. Currently in early development stages, the device could be released as soon as 2027, according to The Information.
OpenAI has introduced an affordable subscription tier called ChatGPT Go, now available globally after its initial launch in India and expansion to over 170 countries. For $8 per month, Go users receive enhanced features compared to the free ChatGPT, including increased message limits, file uploads, and image generation capabilities. Positioned between the free service and the $20 ChatGPT Plus plan, ChatGPT Go targets everyday users for activities such as writing, learning, and creative tasks. The rollout reflects strong user adoption and engagement in markets where Go was previously offered.
Google has broadened access to its AI-driven video creation tool, Flow. Originally launched last May for Google AI Pro and AI Ultra subscribers, Flow is now available to users on Business, Enterprise, and Education Workspace plans. Powered by Google's Veo 3.1 video generation model, Flow creates short, eight-second video clips from text prompts or images. Users can combine these clips into longer scenes and utilize features to adjust lighting, camera angles, and edit objects within the scene. Additionally, Google recently introduced vertical video support to enhance the tool's capabilities.
Google is actively contesting a federal court's ruling that labeled it an illegal monopoly in the online search market. The company submitted an appeal notice on Friday, seeking to suspend the court-mandated measures designed to revive competition in the sector. Lee-Anne Mulholland, Google's VP of regulatory affairs, argued that the court overlooked the fact that users choose Google voluntarily. She emphasized that the decision did not consider the fast-moving innovation and robust competition Google faces from both established companies and emerging startups.
Mike Krieger, Instagram's co-founder and Anthropic's chief product officer since two years ago, is shifting roles to co-lead Anthropic's internal incubator called the "Labs" team. Launched in mid-2024 with only two members, the Labs team is expanding to focus on developing experimental products. Krieger will now serve as a member of technical staff reporting to Anthropic president Daniela Amodei and will co-lead the team alongside Ben Mann, the product engineering lead. Concurrently, Ami Vora, currently the head of product, will assume Krieger's previous responsibilities.
New York Governor Kathy Hochul is prioritizing children's online safety as a key focus of her administration, setting her sights on Roblox, a platform often overlooked in regulatory discussions. With federal progress on internet reforms slow, New York is stepping up with state-level measures to enhance protections for kids online. Hochul unveiled a comprehensive plan to strengthen parental controls and implement age verification, drawing on legislation proposed by state lawmakers Andrew Gounardes and Nily Rozic. This initiative aims to address current regulatory gaps and boost security on platforms popular with young users.
The UK is set to enforce a law criminalizing the creation and commissioning of nonconsensual intimate deepfake images, a measure triggered by the rise of such content on X linked to the Grok AI chatbot. The Data Act, passed last year, designates this act as a criminal offense. Liz Kendall, the UK Secretary of State for Science, Innovation and Technology, announced that this offense will also be prioritized under the Online Safety Act to strengthen protections.
Meta is set to lay off approximately 10% of its Reality Labs team, primarily impacting employees working on metaverse projects. This move reflects the company's strategic pivot toward artificial intelligence, which is drawing resources away from virtual reality initiatives. Andrew Bosworth, Meta's CTO and head of Reality Labs, has called an important meeting for staff to discuss these changes. Reality Labs currently employs about 15,000 people, and the layoffs mark a significant shift in Meta's focus away from the metaverse.
At this year's CES, humanoid robots showcased significant progress toward becoming practical home assistants. LG introduced CLOiD, designed to handle tasks such as cooking and loading laundry machines. Other companies like SwitchBot presented the Onero H1, aimed at simplifying daily chores. Additionally, Boston Dynamics, WIRobotics, Zeroth, and several others revealed cutting-edge humanoid robots. While robots have already mastered tasks like vacuuming and lawn mowing, mastering laundry remains a challenge. Advances in robotics and AI continue to push these capabilities forward, raising the question of how soon robots will fully integrate into our home routines.
