Google is rolling out a beta of Privacy Sandbox for Android starting today. The program is the company’s attempt to blend user privacy with targeted advertising, something the search giant has worked on for years in its planned shift away from cookie-based web tracking.

One of Privacy Sandbox’s pillars is the Topics API, which pulls a list of your top interests based on usage. It then compares them to a database from the Interactive Advertising Bureau and Google’s data. Partner publishers can ping the API, which responds with a list of interests to help serve relevant ads without sharing overly intrusive information. Google says stored interests are “kept for only three weeks, and old topics are deleted.” In addition, the data and processing are done on-device “without involving any external servers, including Google servers.”

The beta is the first time Privacy Sandbox has been available publicly on Android. Google is still working on Privacy Sandbox for Chrome (here’s the timeline), which it has been letting developers test for about a year. It says it received feedback from hundreds of companies, which has helped shape its approach.

Apple

Privacy Sandbox is Google’s answer to Apple’s App Tracking Transparency (ATT), which the iPhone maker introduced in iOS 14.5. The feature requires user consent to track them across other apps and websites. Google dismisses ATT as a “blunt approach” since it doesn’t offer an alternate way for app developers and advertisers to replace the lost income from cookie-based targeting.

Privacy vs. advertising is an arms race. When platforms like iOS block the old ways of profiting from ads, advertisers can (and do) resort to fingerprinting: collecting seemingly innocent device information that, when pieced together, may identify you nearly as well as cookies. Unfortunately, it’s difficult to detect and prevent. Google hopes Privacy Sandbox will strike the right balance between privacy and advertiser / developer revenue.

Google says the Privacy Sandbox beta will roll out gradually, starting with “a small percentage of Android 13 devices” and expanding from there. You’ll see a notification on your device inviting you to join the beta if selected. After accepting, you can visit the new Privacy Sandbox section in your device’s Settings menu to view tracked topics and opt out individually. Additionally, the menu lets you leave the program.

Amazon is moving forward with a live-action Spider-Man Noir series, according toVariety. It will reportedly focus on “an older, grizzled superhero in 1930s New York City” — one that isn’t Peter Parker.

Spider-Man Noir is an alternate version of the web-slinging hero, first seen in the 2009 Marvel comic series of the same name. The comic version was set in 1933, as a freshly bitten Spidey navigates New York City’s Depression-era criminal underworld. The character, who exists in a separate universe from the mainline Spider-Man stories, also appeared in the 2018 film Spider-Man: Into the Spider-Verse, voiced by Nicolas Cage. The report doesn’t mention whether Cage will be involved in the new project, but the actor has said he wasn’t asked to return for the upcoming animated sequel, Spider-Man: Across the Spider-Verse.

Marvel

The as-yet-untitled series will be Amazon’s second project based on Sony-controlled Marvel superheroes after the upcoming Silk: Spider Society. Oren Uziel will write and executive-produce the show; Spider-Man: Into the Spider-Verse alumni Phil Lord, Christopher Miller and Amy Pascal share executive-producing credits.

Sony controls the film rights to Spider-Man and supporting characters like Venom, Carnage, Vulture, Black Cat and others. In addition, it works with Marvel Studios on the current MCU film franchise starring Tom Holland.

1Password is going all-in on passkeys starting this summer, as the company announced today that its users would soon have the option of using passwordless logins. The move follows Google's and Apple’s passkey support starting last year.

Passkeys can replace passwords by providing unique digital keys stored locally on your device. They are much less susceptible to theft in the case of data breaches or phishing attempts and can integrate with your device’s biometric security — including fingerprints and facial logins. 1Password’s passkey support will be “built on the same security foundation” as secret keys, the long recovery codes platforms generate as a fallback sign-in option.

“All you’ll ever need to sign in to 1Password, unlock your vaults, and securely access your data is your one passkey,” the company said in a blog post. “More and more sites and services are adding passkey support every week, but whether you’re first in line to start using them, or you need to rely on passwords for a while longer, we’ve got you covered.”

“Unlike user-created passwords, passkeys are strong and unique by default. They’re generated and stored on your devices, and they’re never shared with our cloud service,” the company said. “Passkeys are also resistant to phishing, and they have a full 256 bits of entropy to prevent cracking — providing even more protection than our Secret Key. They’re safeguarded by biometrics and hardware-level security. And we’re building them to be portable between all your devices and platforms.”

Microsoft announced new AI-powered classroom tools today. The company sees its new “Learning Accelerators” as helping students sharpen their speaking and math skills — while making teachers’ jobs a little easier — as children prepare for an even more technologically enhanced world.

Speaker Progress is a new AI classroom tool for teachers. Microsoft says it saves them time by “streamlining the process of creating, reviewing, and analyzing speaking and presentation assignments for students, groups, and classrooms.” It can provide tidy summaries of presentation-based skills while highlighting areas to improve. Additionally, it lets teachers review student recordings, identify their needs and track progress.

It will be a companion for Speaker Coach, an existing feature Microsoft launched in 2021 that provides one-on-one speaking guidance and feedback. For example, it uses AI to give real-time pointers on pacing, pitch and filler words. “Speaker Coach is one of those tools that kind of was a lightbulb tool for a lot of students that I’ve worked with,” said an unnamed teacher in a Microsoft launch video. “Being able to practice and get real-time feedback is where Speaker Coach really comes in and helps our students, and it even helps us as adults.”

Microsoft

Microsoft’s AI math tools are its answer to nosediving math scores during the pandemic. Math Coach deconstructs problems, walking students through the steps to solve them while encouraging critical thinking. Meanwhile, Math Progress is the teacher-focused companion tool, helping them generate practice questions and provide more efficient feedback. The company says the features work together: Math Coach uses teacher input from Math Progress to develop new lessons. Additionally, it says schools can use the tools’ overall math fluency data to track progress and better meet their goals.

Speaker Progress, Math Coach and Math Progress will launch in Microsoft Teams for Education in the 2023-24 school year. Meanwhile, Speaker Coach is available now in Teams and PowerPoint.

TP-Link briefly teased its entry into robot vacuums at CES, but today the company released more detail on its first pair of budget-minded cleaning machines. The Tapo RV10 is a $250 vacuum / mop combo, while the Tapo RV10 Lite cuts the mop, doubles the dustbin size and saves you $20.

Both models use gyroscope-based “zig-zag path planning,” which the company says will clean your floor more efficiently than random-path vacuums. In addition, the machines offer four suction levels (quiet, standard, turbo and max) with a maximum of 2,000Pa suction power. They also include noise suppression to keep the sound as low as 53dB in quiet mode. Both devices work with the Tapo app for iOS and Android; Alexa and Google Assistant voice control are also supported.

TP-Link

The mopping RV10 differentiates itself from its vacuum-only sibling with a 300ml water tank, an electronic pump and three water flow levels. Additionally, the RV10 only has a 400ml dustbin, presumably to make room for its mopping hardware; the RV10 Lite’s bin doubles the size at 800ml.

The vacuums each have a 2,600mAh battery, adding up to an estimated three hours of cleaning per charge. TP-Link says if their batteries get low before finishing, they’ll dock automatically, recharge and pick up where they left off.

The vacuums are available today on Amazon. The Tapo RV10 costs $250, while the Tapo RV10 Lite is $230.

Spotify is launching a new feature today called Exclude from your Taste Profile. As its name suggests, it lets you choose playlists — perhaps those played by a child or other family member — that you don’t want affecting your personalized recommendations.

The company says the feature “reduces the impact” the playlists have on your recommended content. So if your child loves jamming out to “Elmo’s Favorite Elmo Songs!” ad nauseam, the new feature should help you avoid hearing similar children’s tunes in your Daily Mix playlists and other algorithmically generated content. (Although we wouldn’t fault you for putting on “Brushy Brush!” while cleaning those pearly whites.)

Exclude from your Taste Profile only works with playlists — not albums or individual songs. Spotify says liked songs within those playlists will still affect your recommendations, so it may be wise to ask family members to ease up on the heart button. The changes will apply retroactively, excluding past and future listens of the unwanted playlists from your recommendations.

Spotify is rolling out the feature “starting today,” suggesting you may not see it immediately. The feature will work on iOS, Android, desktop and the web. Once available, you can exclude music by selecting a playlist, tapping the three-dot menu near the top and selecting “Exclude from your Taste Profile.” In addition, the company says you can turn the feature off at any time using the same process.

If the unofficial debut of Google’s Bard chatbot is any indication, misinformation is about to get a lot worse. The company posted an ad to Twitter this week showing off the natural-language AI model displaying false information about the James Webb Space Telescope (JWST).

In the advertisement (via Reuters), a short GIF shows an example of a Q&A with Bard. “What new discoveries from the James Webb Space Telescope can I tell my 9-year old about?” the query reads. The machine quickly spits out three ideas, including the last one that says, “JWST took the very first pictures of a planet outside of our own solar system. These distant worlds are called ‘exoplanets.’ Exo means ‘from outside.’” Although the bit about exoplanets is spot-on, the first part saying the James Webb Space Telescope (JWST) took the first pictures of them is false. That honor belongs to the European Southern Observatory’s Very Large Telescope (VLT) in 2004, as confirmed by NASA.

Although incorrect information in a Twitter ad won’t likely hurt anything directly, it’s easy to view the mistake as an omen of the risks of releasing natural-language chatbots into the wild. It parallels CNET’s decision to write financial advice articles with an AI chatbot; they were also riddled with errors.

Because chatbots get so much right — and spit out answers with such supreme confidence — anyone who doesn’t fact-check their responses may be left with false beliefs. Considering the chaos that (non-AI-powered) misinformation has already let loose on society, releasing this often mind-blowing technology before it can be trusted to produce factual information reliably and accurately — even escaping Google’s copy editors — we may be in for a wild ride.

Hot Wheels: Rift Rally is a new mixed reality game from Velan Studios, the creators of Mario Kart Live: Home Circuit. Like Nintendo’s 2020 title, you control a real RC car while your screen shows an augmented view from its cockpit. It will be available in March for PS5 / PS4 and iOS.

The game bundle includes one toy car, which you can drive in two modes: Campaign and Stunt. In Campaign, you set up gates around your room, which you drive the car through to complete the different maps (each includes unique challenges). Meanwhile, Stunt mode skips the gates, letting you drive wherever you want while performing digitally enhanced drifts, wheelies and burnouts to beat your high scores.

It supports a solo campaign and multiplayer. When playing with a friend, they can bring their car for competitive racing or (on PlayStation only) play co-op using a single toy vehicle as the action switches between controllers.

Velan Studios

The bundled Chameleon car can appear onscreen as any of 22 different virtual vehicles with over 140 variants. It includes “endless options” for digital customization, including performance mods that change how the car moves across your space; for example, adding an acceleration boost in-game will make the toy car drive faster. In addition, you can play various mini-games, puzzles and challenges to unlock new vehicles and mods.

Although the software is free to download, you’ll need to buy at least a $130 Standard Edition, which gets you the Chameleon car, four Rift gates and a charging cable. A $150 Collector’s Edition swaps in a “special edition” Chameleon while adding a (non-playable) McLaren Senna Hot Wheels die-cast toy. Hot Wheels: Rift Rally launches on March 14 and is available for pre-order starting today.

A Minecraft mad scientist has recreated The Legend of Zelda inside the blocky sandbox game without any third-party mods or resource packs. Fan and YouTuber C1OUS3R, who has also made Sonic the Hedgehog and Super Mario Bros. inside Minecraft, took nearly a month to craft the Zelda tribute.

The developer used command blocks, custom textures and other in-game features to recreate the 1986 NES classic. “I use Blockbench to create a flat version of Link from all sides so I can rotate it to make it look like he’s facing every direction,” C1OUS3R explained in a making-of video. “I would make a Voxel model like I did in the previous video games, however, it’s pretty much unnoticeable and takes 30 times longer, and I don’t really have the time for that.

“I then overlaid this model over a carved pumpkin which is one of the items you can wear on your head in Minecraft. I then make the player invisible to make it look like you’re controlling Link. Then by using the data packs function folder, which is just command blocks in written form, I’m able to detect which way the player is moving based off armor stands and rotate the model to make it look like it’s moving. I then simply animate the textures, and boom. You have a controllable Link.”

C1OUS3R said on Reddit (viaKotaku) they will release the playable creation, which runs through the first boss fight, once their YouTube video gets 5,000 likes. (It has over 500 at the time of publication.) The developer says they want to eventually release an entire gaming engine inside Minecraft. “It’s much easier to understand rather than something like Unity. I believe Minecraft has the ability to really help people get into game development.”

Twitch has banned “Nothing, Forever,” the AI-generated Seinfeld stream, for at least 14 days following a transphobic and homophobic outburst. It’s the latest example of “hate in, hate out” when AI chatbots are trained on offensive content without adequate moderation.

Like Seinfeld, “Nothing, Forever” rotates between standup bits and scenes in the comedian’s apartment (he’s called “Larry Feinberg” in the AI version). As first reported by Vice, during one of the recent AI-scripted standup acts, the Seinfeld counterpart suggested that being transgender is a mental illness. In what almost seemed like an awareness of the material’s offensiveness, the AI comedian quickly added, “But no one is laughing, so I’m going to stop. Thanks for coming out tonight. See you next time. Where’d everybody go?” 

Although Twitch hasn’t confirmed that the “joke” was the reason for the ban, the stream was removed soon after the problematic segment aired. The program’s creators blame the hurtful rant on a model change that inadvertently left the stream without moderation tools.

“Earlier tonight, we started having an outage using OpenAI’s GPT-3 Davinci model, which caused the show to exhibit errant behaviors (you may have seen empty rooms cycling through),” a staff member wrote on Discord. “OpenAI has a less sophisticated model, Curie, that was the predecessor to Davinci. When davinci started failing, we switched over to Curie to try to keep the show running without any downtime. The switch to Curie was what resulted in the inappropriate text being generated. We leverage OpenAI’s content moderation tools, which have worked thus far for the Davinci model, but were not successful with Curie. We’ve been able to identify the root cause of our issue with the Davinci model, and will not be using Curie as a fallback in the future. We hope this sheds a little light on how this happened.”

Twitch

The team elaborated in another Discord post (viaThe Verge). “We mistakenly believed that we were leveraging OpenAI’s content moderation system for their text generation models. We are working now to implement OpenAI’s content moderation API (it’s a tool we can use to verify the safeness of the content) before we go live again, and investigating secondary content moderation systems as redundancies.”

Although the team sounds genuinely apologetic, stressing that the bigoted rant was a technical error that doesn’t represent their views, it reiterates the importance of consistent AI moderation. You may remember Microsoft’s Twitter chatbot, which only lasted about 16 hours after users taught it to spew conspiracy theories, racist views and misogynistic remarks. Then there was the bot trained entirely on 4chan, which turned out exactly as you’d expect. Whether “Nothing, Forever” returns or not, the next time a team of developers is faced with a choice between unexpected downtime and making sure those filters are in place, pick the latter.