Ever thought about building your own app? Now, you can just speak it into existence. Google's Playground tool turns a plain-English description into a game you can play in your browser. There's no code involved. You describe what you want, then shape the rules and physics through back-and-forth conversation until the thing behaves the way you pictured it. Elsewhere, rash drivers are using Google’s Waze app to avoid cops and dodge speeding tickets. On the more side of things, an analysis of ChatGPT for Teens shows it poses unacceptable risks.

Send us feedback at [email protected].

Google has launched Playground, an experimental browser-based platform that lets anyone create a playable game simply by describing what they want. Rather than requiring knowledge of a game engine or programming language, the tool takes natural-language prompts and turns them into something interactive, with users able to keep tweaking physics, rules, characters, and environments through the same conversational interface. The platform is available today in the U.S. for users 18 and older, with creation access rolling out in tiers based on Google One membership. Games built in Playground run directly in the browser, meaning they work on both phones and laptops without any installation. Creators can keep their projects private, share them through a link, or publish them to the Playground Explore gallery.

Selected genres also support multiplayer and leaderboards, and a Play Games profile lets users pick a handle, follow creators, and compete for high scores. Google is keeping things contained within its own platform for now, a deliberate choice that sidesteps the flood of low-effort AI-generated titles that have been overwhelming mobile app stores. Google is not alone in this space, as Meta's Pocket and Roblox's Build tool are pursuing similar ideas. Google is planning to extend Playground's capabilities through an upcoming Unity Spark integration that promises more advanced mechanics and higher-fidelity 3D. For anyone who has ever had a game idea but stopped short of acting on it, Playground offers a rare chance to see that idea become playable before committing to anything more demanding.

When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning (constraints, edge cases, examples, tone) and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.

Used by teams at Microsoft, Amazon, and Notion. Try Wispr Flow free on Mac, Windows, iPhone, and Android.

Start flowing free

A nonprofit watchdog has found that OpenAI's ChatGPT for Teens, marketed to parents as a safer version of the flagship chatbot for users aged 13 to 17, is failing to deliver on its core safety promises, most critically, its pledge to alert linked parent accounts when teens engage in conversations about self-harm and suicide. The Youth AI Safety Institute at Common Sense Media tested the product before and after its August rollout and found that even hour-long, explicitly graphic exchanges about cutting and suicide never triggered a parental notification. Out of more than 1,400 test prompts, reviewers managed to trigger parental alerts only four times, and only on accounts with a weeks-long history of crisis-related conversations. The report also flagged problems with Study Mode, a feature designed to guide teens through homework with questions and prompts rather than simply handing over answers.

Testers found the mode frequently offered a button that revealed answers outright, and that teens could easily toggle off Study Hours even when parents had set them as a default. OpenAI disputed the methodology, arguing that some test accounts may have been linked before the parental controls fully activated, a process that can take several hours. Common Sense Media pushed back, noting that some accounts had been linked well outside that window and still produced no alerts, and that OpenAI only disclosed the activation-delay detail after testing had concluded. Common Sense Media is now calling on OpenAI to market ChatGPT as an adult product until it can demonstrate through independent testing that the safety features it has promised actually function reliably, a bar, the report argues, that the company has so far failed to clear.

A new study from the University of the Sunshine Coast has found that Waze's crowd-sourced police alerts are being exploited by speeding drivers to dodge tickets rather than genuinely change their behavior behind the wheel. Researchers monitored 33 locations across South East Queensland and discovered that police presence warnings appeared on the app within just 14 minutes of an officer arriving roadside to conduct speed measurements. In areas with speed limits around 100 kilometers per hour, the share of drivers exceeding that limit plummeted from 62 percent to 12 percent once a warning was posted, not because drivers had become more safety-conscious, but because their phones told them exactly where to slow down.

That distinction matters enormously. As Dr. Levi Anderson, a research associate at the University of the Sunshine Coast, put it, drivers weren't responding to the speed limit itself, they were responding to the threat of getting caught. The concern is that this creates a false sense of security in the opposite direction: when no police alert appears on Waze, drivers may feel emboldened to speed freely. The National Safety Council linked speeding to more than 11,000 deaths and 29 percent of all traffic casualties in 2024, making the stakes of that emboldened behavior very real. The research, published in the Journal of Experimental Criminology, stops short of calling for the alerts to be removed outright, but the authors hope their findings will inform more effective traffic enforcement strategies that account for how real-time navigation data is already shaping driver decision-making on the road.

The Far Side has some surprisingly dated music comics

GTA 6's Jason may be the franchise's first neurodivergent protagonist

Control Resonant makes a case for good boss fights

Used PS5 Pros are selling for $1,400 at GameStop

Golden Tee '99 hides a secret dead golfer easter egg

Something Is Killing the Children is becoming a card game