OpenAI has added age prediction features to ChatGPT, meant to identify and bolster protections for underage users. These age detection plans were announced in December alongside updated guidelines for interacting with teens, and follow a rise in similar age-gating efforts from online platforms, including Instagram, YouTube, TikTok, and Roblox. ChatGPT's age prediction model works by examining behavioral and account-level signals, including a user's stated age, how old the account is, when the user is active, and usage patterns over time.
In 2016, I presented at @Roblox Indie Game Developer Meetup about design strategy as an indie developer. Back then, I had no idea children as young as 5 were interacting with random adults on their platform. Today, the same company (NYSE: $RBLX) is filled with poorly moderated "games" like Bathroom Simulator and worse - all while letting adults animate their avatars for sexual role play.
Modern digital teams depend on reliable data flows - whether that's monitoring search visibility, validating ads, testing apps across regions, or running automated QA at scale. In all of these cases, a proxy layer can help businesses manage IP rotation, session control, and traffic distribution in a way that's more predictable than relying on "best effort" network conditions. This review looks at Proxys.io from a practical, business-first angle:
Google this week announced the public rollout of a beta Gemini feature called Personal Intelligence. It gives Gemini access to your Gmail, Google Photos, YouTube, and Google Search data, if you grant permission, which you can do individually (for example, you can grant permission for YouTube access but not Gmail). The rollout will take place over the coming weeks, and the feature is exclusive to English-speaking Google AI Pro or AI Ultra plan holders in the U.S.
If you track your children's location using tech, you're not alone. A 2024 survey by Pew Research Center found that one in four parents monitor their kids' whereabouts using GPS. But is that private data safe? Consumer Reports evaluated 15 popular kid-tracking devices, and here's what they found. Overall, Apple's AirTags and Apple Watches, eufy's SmartTrack Link, and Garmin's Bounce performed well when it came to privacy and data security.
Poland's power system faced its largest cyberattack in years in the last week of December that also followed a different pattern, the country's energy minister said on Tuesday.
Authorities have responded with severe measures, including a nationwide telecoms blackout and jamming satellite services like Starlink, aimed at preventing coordination among demonstrators. Iranians are embracing freedom tech tools; Bitchat, Noghteha, and Delta Chat for offline communication. Two of these apps trace their origins directly to Bitcoin, highlighting how technologies from this community provide practical solutions in high-stakes environments. Bitchat, built by Bitcoin pioneers Jack Dorsey and open-source developer Calle, operates over Bluetooth mesh networks and the Nostr protocol without needing an internet connection.
Over the past five years, annual robocall volume has consistently remained between approximately 50 billion and 55 billion, according to the YouMail Robocall Index. Robocall volume for 2025 totaled 52.5 billion, down a little over 1% from the 2024 total of 52.8 billion. December's 4.1 billion robocalls were up 6.4% from November but down 5.7% from December 2024. YouMail noted that robocalls increased in November and December 2025 after reaching a multi-year low in October.
The conservative writer and political strategist Ashley St Clair had just put her young son to bed when she received a text from a friend with a link to a page on X. When she opened it, she says she saw a photo of Grok undressing me and putting me in a bikini'. St Clair was not alone. In recent weeks, countless women have found their images digitally manipulated online by Grok, at the request of X users, to disrobe them.
Our pick for best premium VPN is fast and easy to use. ExpressVPN is back on sale again, and its two-year plans are up to 78 percent off right now. You can get the Advanced tier for $101 for 28 months. This is marked down from the $382 that this time frame normally costs. On a per-month basis, it works out to roughly $3.59 for the promo period.
ICE has been invading U.S. cities, targeting, surveilling , assaulting , and people who are undocumented immigrants. They also have targeted people with work permits , asylum seekers , permanent residents (people holding "green cards"), naturalized citizens , and even citizens by birth. ICE has spent hundreds of millions of dollars on surveillance technology to spy on anyone - and potentially everyone-in the United States.
On smart TVs, ACR can capture hundreds of images of what's on your screen each minute, whether that content originates from the TV's built-in apps or a connected device like a dedicated streaming box or game console. The screenshots are sent to your TV's manufacturer, who can use the data gathered to target ads on your TV, or sell the data to other ad providers to target ads on other platforms.
Turns out anyone with an internet connection, and not even a password, could access the feeds of a line of the Flock surveillance cameras that are currently all the rage with the SFPD and Oakland Police Department. Below we see pictured what is likely one of those Flock Safety automated license plate reader cameras, one of the 400 of these new security cameras in SF that are are taking three million surveillance photos every day.
Considering how much personal information we store on our smartphones, I wouldn't be surprised if the demand for privacy screen protectors is at an all-time high. From bank statements to email credentials to photos and videos, the need to keep our digital footprint safe and secure, especially from the prying eyes of strangers, has never been more necessary. Also: Stop putting your phone screen-side up on the table - why you'll be glad you did Earlier this year, I gave the privacy-enhanced accessory a try, installing on my Galaxy S25 Ultra. The installation process was the same as any other modern-day screen protector, with a guide tray that you set your Samsung phone into for near-perfect placement.
If you click 'Technical identifiers Accept all', we and our partners, including 245 who are part of the IAB Transparency & Consent Framework, will also store and/or access information on a device (in other words, use cookies) and use precise geolocation data and other personal data such as Technical identifiers are system-generated strings of letters and numbers that can identify your device or you as a user.
Most phones make a familiar bargain: free services and slick apps in exchange for constant tracking, profiling, and data being treated as currency. The line about how if you do not pay for the product, you are the product, has gone from cliché to lived reality. Punkt. has been quietly pushing back against that logic for years, starting with minimalist feature phones and now moving into full touchscreen territory with the same philosophy intact.
California is giving residents a new tool that should make it easier for them to limit data brokers' ability to store and sell their personal information. While state residents have had the right to demand that a company stop collecting and selling their data since 2020, doing so required a laborious process of opting out with each individual company. The Delete Act, passed in 2023, was supposed to simplify things, allowing residents to make a single request that more than 500 registered data brokers delete their information.
Contextual integrity defines privacy as the appropriateness of information flows within specific social contexts, that is, disclosing only the information strictly necessary to carry through a given task, such as booking a medical appointment. According to Microsoft's researchers, today's LLMs lack this kind of contextual awareness and can potentially disclose sensitive information, thereby undermining user trust. The first approach focuses on inference-time checks, i.e., safeguards applied when a model generates its response.
