Microsoft wants to make generative AI a big part of Windows and the PCs that run it.

This week, at a keynote event before its annual Build developer conference, the company showed off a new line of Windows computers it calls Copilot+ PCs. The computers also have generative AI-powered features like Recall that help users find apps, files, and other content they’ve seen before. Copilot is Microsoft’s brand of generative AI, and it will soon be a lot more built into Windows 11. Also, new Microsoft Surface gadgets are on the way.

All the important news has been put together here.

The Copilot+ PCs

Copilot+ PCs are Microsoft’s idea of the best Windows hardware for AI. All of them have special chips, called NPUs, that power AI games like Recall. They also come with at least 16GB of RAM and SSD storage.

Qualcomm’s Snapdragon X Elite and Plus chips will be in the first Copilot+ PCs. Microsoft says these chips can power up to 15 hours of web browsing and 20 hours of video playback. Intel and AMD have also promised to work with a number of manufacturers, such as Acer, Asus, Dell, HP, Lenovo, and Samsung, to make processors for Copilot+ devices.

Some Copilot+ PCs can be preordered today for as little as $999.

Surface Book and Surface Pro
The Surface Laptop and Surface Pro, which Microsoft just released, are designed to be fast and last a long time.

The most recent iteration of the Surface Laptop, which comes in either a 13.8- or 15-inch size, has undergone a redesign featuring sleeker screen bezels and contemporary aesthetics. According to the company, the device has a battery life of up to 22 hours and is claimed to be 86% faster than the Surface Laptop 5. Additionally, it is compatible with Wi-Fi 7 and features a touchpad with haptic feedback.

Regarding the new Surface Pro, Microsoft claims that it offers a performance boost of up to 90% compared to the previous-generation Surface Pro (the Surface Pro 9). It also features a new OLED display with HDR, Wi-Fi 7 (with the option for 5G connectivity), and an improved ultrawide front-facing camera. In addition, the keyboard of the device has been strengthened with extra carbon fiber and now includes haptic feedback.

Retrieve

Windows 11’s upcoming Recall feature has the ability to retain information about apps and content that a user accessed on their PC, even if it was weeks or months ago. This feature can be helpful in locating specific information, such as in a Discord chat where the user was discussing potential clothing purchases. Recall’s timeline feature allows users to easily review their recent work and navigate through files, such as PowerPoint presentations, to find relevant information for their searches.

Microsoft claims that Recall has the capability to establish connections between colors, images, and other elements, enabling users to search for virtually anything on their personal computers using natural language, similar to the technology employed by startup Rewind. The company asserts that all user data linked to Recall is maintained as private and stored on the device without being utilized for training AI models, which is of significant importance.

Here is additional information from Microsoft: “Your snapshots are stored exclusively on your personal computer and are not shared or stored elsewhere.” To delete specific snapshots, modify and remove time intervals, or pause the process, you can access the corresponding options in the Settings menu. Alternatively, you can directly perform these actions by clicking on the icon located in the System Tray on your Taskbar. You also have the option to apply filters that prevent certain apps and websites from being saved.

Photo manipulation and real-time language translations
Windows now incorporates a greater amount of artificial intelligence (AI) than ever before, with certain AI features available exclusively on the new Copilot+ PCs.

The Super Resolution feature has the ability to automatically enhance and enlarge old photos. Copilot now has the capability to analyze images and provide users with suggestions for creative compositions. Users can utilize the Cocreator feature to generate images and instruct the AI model to modify or transform the image based on their drawings.

Additionally, live captions with live translations can translate any audio that passes through a PC, regardless of its source (such as YouTube or a local file), into the user’s preferred language. Live translations will initially support approximately 40 languages, including English, Spanish, Mandarin, and Russian.

The Windows Copilot Runtime is a crucial component for your system.
Enabling features like Recall and Super Resolution is the Windows Copilot Runtime, a set of approximately 40 generative AI models that Microsoft refers to as “a new layer” of Windows. Similar to an IT project manager, the Windows Copilot Runtime enables generative AI-powered apps, including third-party apps, to run on an individual Copilot+ PC without relying on an internet connection. The semantic index and a vector-based system are responsible for making this possible.

According to Microsoft’s announcement, CapCut, the well-known video editor from TikTok owner ByteDance, will use the Windows Copilot Runtime to improve the effectiveness of its AI capabilities.

OpenAI is taking away one of the voices that ChatGPT uses. The company announced on Monday that users thought it sounded like Scarlett Johansson. In response, Johansson said in a statement that she had hired a lawyer to look into the Sky voice and find out exactly how it was made. The game Sky is being stopped because OpenAI used it last week to show off its new GPT-4o model.

According to the company in a blog post, AI voices shouldn’t try to sound like famous people. For example, Sky’s voice is not an imitation of Scarlett Johansson’s; it belongs to a different professional actress who uses her own natural speaking voice. “To protect their privacy, we can’t say who does our voice work.”

Last week, a video of the demo went viral on social media because people thought the voice sounded like Scarlett Johansson’s. There were jokes about how flirtatious the voice was, and some people compared it to a man’s fantasy.

Say hello to GPT-4o, our new flagship model which can reason across audio, vision, and text in real time: https://t.co/MYHZB79UqN

Text and image input rolling out today in API and ChatGPT with voice and video in the coming weeks. pic.twitter.com/uuthKZyzYx

— OpenAI (@OpenAI) May 13, 2024

People have said that the flirty voice sounds a lot like Scarlett Johansson’s role as a seductive virtual assistant in the 2013 movie “Her.” Joaquin Phoenix plays the lead role in the movie, and the main character falls in love with the virtual assistant.

The company hasn’t said anything about the similarity between Sky’s voice and Johansson’s, but OpenAI CEO Sam Altman tweeted the word “Her” after the event.

her

— Sam Altman (@sama) May 13, 2024

OpenAI’s demo last week was supposed to show off the chatbot’s better ability to have conversations, but it went viral when the sultry voice laughed at almost everything an OpenAI employee said. It said to the worker, “Wow, that’s quite the outfit you’re wearing at one point.” There was another time when the chatbot got complimented and said, “Stop it, you’re making me blush.”

It says in a blog post that it wants the voices of its chatbots to sound “approachable” and “inspire trust.” It also wants them to have a voice that is “warm, engaging, confidence-building, and charismatic.”

In the future, OpenAI says it will “add more voices to ChatGPT to better match the wide range of users’ interests and preferences.”

The whole statement from Johansson is:

“In September, Sam Altman asked me to work for him as a voice actor for the current ChatGPT 4.0 system.” He told me that he thought that by speaking out about the system, I could help tech companies and artists work together and make people feel better about the huge change between humans and AI. He told me that he thought my voice would make people feel better.

I turned down the offer after giving it a lot of thought and for personal reasons. After nine months, everyone, including my friends, family, and strangers, said that the new system called “Sky” sounded a lot like me.

When I heard the demo that was made public, I was shocked, angry, and confused that Mr. Altman would be interested in a voice that sounded so much like mine that even my closest friends and news outlets couldn’t tell the difference. By tweeting the word “her,” Mr. Altman even made it sound like the resemblance was on purpose, referring to the movie in which I voiced Samantha, a chatbot who becomes close with a human.

Mr. Altman called my agent two days before the ChatGPT 4.0 demo came out and asked me to think again. It was out there before we could connect.

Because of what they did, I had to hire a lawyer. The lawyer sent two letters to Mr. Altman and OpenAI explaining what they had done and asking them to explain in detail how they made the “Sky” voice. Because of this, OpenAI reluctantly agreed to remove the “Sky” voice.

Many of us are dealing with deep fakes and how to protect our own likenesses, work, and identities right now. I think these are questions that need to be answered completely. “I look forward to resolution in the form of openness and the passing of appropriate legislation to help protect individual rights.”

OpenAI shared this quote from Altman: “Sky’s voice is not Scarlett Johansson’s, and it was never meant to sound like hers.” Before we reached out to Ms. Johansson, we hired the voice actor who would be Sky’s voice. Because we respect Ms. Johansson, we have stopped using Sky’s voice in our apps and games. We’re sorry that we didn’t talk to Ms. Johansson better.

Imagine yourself engaged in your own affairs, seated on a park bench, gazing at your mobile device. Explosion. An individual seizes your device and swiftly flees with it. While Android and iOS devices do have certain security measures, what about the brief period of time when the phone is still unlocked? Is there a method available to remotely erase its data?

Burglars can obtain a substantial amount of information within that brief duration. Each moment is significant. During the Google I/O 2024 developer conference, Google unveiled a new feature for Android called Theft Detection Lock. This feature is specifically designed to safeguard against the increasing risk of theft. Once activated, the AI-driven function will automatically secure the device.

According to Google, if your phone detects a typical movement related to theft, it will rapidly lock the screen to prevent thieves from easily accessing your data. An instance of such a stimulus is a mechanism that abruptly initiates rapid motion in the opposite direction.

Google is implementing an offline device lock feature, specifically designed to safeguard the device in the event of intentional disconnection from the network. Occurrences such as consistently failing to authenticate the phone will activate that functionality.

The forthcoming update will also introduce functionality that enhances the level of difficulty for malefactors attempting to perform a remote factory reset on your device. According to Google, this upgrade prevents thieves from setting up a stolen device again without having knowledge of your device or Google account credentials, even if they force a reset. By rendering a stolen device unsellable, it diminishes the motivation for individuals to engage in phone theft.

Biometric authentication will be mandatory for modifying sensitive information while the device is connected from an unsecured location.