Neowin

Google is deprecating Duplex on the Web, a set of tools that gave Google Assistant the ability to buy movie tickets, check in to flights, change passwords, and track discounts on the user's behalf.

The Telecom Regulatory Authority of India is working to limit text message spam and financial frauds that occur over phone using artificial intelligence and machine learning models.

The new partnership with iCAD makes for Google's first commercial agreement to license its AI-based mammography model and eventually integrate it with real-world clinical products.

Meta announced a network timing protocol for its data centers powering its technologies, Precision Time Protocol. The company claims that PTP protocol can sync time down to nanoseconds.

The U.S. Department of Defense will be benefitting in four key areas as a result of this partnership, including cloud innovation, artificial intelligence, machine learning, and more.

Intel has unveiled the Intel Max Series product family for high-performance computing (HPC) and artificial intelligence ahead of Supercomputing '22 (SC22) in Dallas happening between November 14-17.

At Zoomtopia 2022, Zoom unveiled several new features for the platform like the beta version of Zoom Mail and Calendar, a virtual IQ Coach, a virtual agent, and a hybrid co-working space.

Researchers at MIT have developed a fascinating machine learning model that leverages sound to accurately map an environment and also determine the acoustics of the room at different locations.

Alter, previously known as Facemoji, has now been acquired by Google for $100 million according to an insider. The acquisition happened around 2 months ago without any official announcements.

Shutterstock has announced a new tool for users that allow for the sale of images generated through OpenAI's tools. Dubbed shutterstock.ai, this fund will fund creators and help with training the AI.

Microsoft has detailed a recent effort that makes Teams calls on Windows sound better in challenging network conditions by synthesizing audio through an AI model. It's being tested on Mac too.

The European Union is set to announce the AI Liability Directive that will make it easier for individuals to sue the manufacturer of an AI product or its user through "presumption of causality".

Ahead of COP27, Microsoft, along with Planet and The Nature Conservancy, will launch the Global Renewables Watch designed to map renewable energy installs using AI and satellite images.

Microsoft is working on a nifty new capability for Edge that will allow its browser to automatically enhance images by improving their lighting, contrast, and color. It's currently in preview.

Meta has announced that its management of the PyTorch framework will be handed over to the new PyTorch Foundation, which is in turn a part of the Linux Foundation. A board will guide its future.

Google is rolling out an "AI Test Kitchen" app that lets users test experimental AI-powered projects undergoing development. Currently, the feature is available to some in the US only.

A member of NVIDIA Metropolis, Seoul Robotics, is making cars autonomous by employing an AI-enabled computer vision platform that is backed by NVIDIA CUDA libraries, along with NVIDIA Jetson AGX Orin.

Xiaomi has introduced its first-ever humanoid robot, CyberOne. It can allegedly detect human emotion and has a self-developed algorithm controlling its many different capabilities.

Xiaomi announced its plans to invest RMB 3.3 billion ($490 million) in developing Xiaomi Pilot Technology. The investment pays R&D costs of the 500 researchers and acquisitions of 10+ enterprises.

NVIDIA reported $2.04 billion in gaming revenue for the first half of 2022, down from the $3.06 billion it made in 2021. AMD, on the other hand, managed to increase its revenue from gaming.

A Google ex-employee claimed that the company's LaMDA AI had become sentient. However, Google did not agree and now, another expert agrees saying that the stories were "pure clickbait".

Meta has launched BlenderBot 3, a new conversational AI. A web demo featuring the chatbot is accessible in the U.S. It said data that it collects will be shared with the scientific community.

This eBook explores how Autonomous Response uses AI's understanding of self to take targeted action against ongoing attacks at every stage and stop in-progress attacks, without disruption.

Google Maps has introduced three features for travel buffs including a 3D immersive view of buildings, routing info for cyclists, and a coordination tool for time-friendly gatherings and meetings.

A robot recently broke the finger of a seven-year-old during a chess tournament in Russia. The machine practically pounced on the child's finger and grabbed it with force, proceeding to break it.

A lot happened in the world of Microsoft this week, ranging from a bunch of Windows builds to leaks about canceled hardware. There was also a fair bit of news on the software side, read about it here!

HeHealth, a Singaporean startup, has come up with a new idea: send an entirely solicited dick pic to the AI for it to check whether the user has a sexually transmitted disease or other problem.

DALL·E, the AI-powered image generator, is entering closed beta, giving access to a million more people. Users will also have commercial selling and merchandising rights for any images they create.

The UK's Ministry of Defence is creating the Defence Centre for AI Research to figure out how it can use artificial intelligence ethically in defence contexts. It also hopes it boosts the UK economy.

Meta has built an language model that translates across 200 languages better than the competition at an average of 44%. It has up to 54.5 billion parameters and has been open-sourced on GitHub.

Riot Games has provided a launch date for a controversial change that will allow it to record your voice chats to train its language model and take action upon reported incidents of hate speech.

Microsoft has announced that it is open-sourcing its latest GODEL conversational AI, which enables goal-oriented interactions while still providing external information, just like a real conversation.

Amazon has demoed an upcoming capability that will enable Alexa to mimic the voice of your deceased relative while conversing with you. It needs just one minute of audio input to work.

Microsoft has announced the general availability of GitHub Copilot, its AI-powered development assistant. If you're looking for an AI to help you write code, you can give it a go at $10/month.

You can now change your voice into Morgan Freeman, a pilot, an astronaut, and several others using this new feature by Voicemod. It is currently accessible on the beta version for PC users.

Samsung's and NAVER's partnership will see the former's network solution supporting the latter's private 5G commercial service to power cloud-based autonomous mobile robots within this month.