logo
Update Every App On Your Phone That's On This List

Update Every App On Your Phone That's On This List

Forbes9 hours ago

Update these apps now.
The recent warning that Meta and Yandex have been secretly tracking billions of phones is a stark reminder that your most sensitive data is at risk. That loophole will now close, as others will be found. Let's not forget Google itself was caught doing broadly the same.
It's now four years since Apple's game-changing App Privacy Labels exposed the sheer extent of data harvesting targeting iPhone users, with the assumption that Android must be even worse. I covered that extensively at the time, and it was clear then — as it is now — that when you're not paying for a product, you are the product.
Multiple reports since have highlighted that permission abuse is still rife, with apps requesting access to data and functions they do not need to deliver the features of the app itself. This is data monetization, pure and simple, your data monetization.
Top-1o data hungry apps
Now the researchers at Apteco have revisited Apple's privacy labels to find out 'who's collecting most of your data' in 2025. The study focused specifically on 'Data linked to you,' as this is the type of data that ties directly back to your identity.'
Apteco's key findings are unsurprising: 'Social media apps are the most data hungry,' and the most collected personal data is 'contact information (such as your name, phone number and home address).' But the range of harvested data goes far beyond that, as you can see from Apteco's table reporting the data accessed during testing.
Data collected from users.
Apteco's list 'is dominated by social media… highlighting how important data collection is to these types of platforms in order to customize content to show things such as posts and friend suggestions [and] which build detailed social profiles.'
Apteco's top-1o list is dominated by global brands with apps installed by hundreds of millions if not billions of users. This isn't a call to delete those apps — albeit you should be aware of the data they're collecting while running on your phone.
App settings on iPhone
Instead, you should update the permissions granted to apps on this list, deciding if you want to give them blanket access to location and other sensitive data. You should also be aware that when you operate within the confines of an app, for example using its own browser, you are not protected by the usual web tracking defenses on your phone.
You don't need to grant all the permissions requested, and you can limit those permissions that might be needed — such as location — to only apply when using the app or to manually request each time before sharing. You can also restrict location data such that it's not precise and just gives a general idea of where you are.
Here are instructions for iPhone and Android on how to apply updates.
'The study highlights how extensive data collection has become across a huge variety of apps,' Apteco says. 'The sheer scale of data collected highlights why understanding and managing app permissions and data policies is increasingly important for users [who] need to be aware of how to actively manage app permissions and data policies.'

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Multimodal AI: A Powerful Leap With Complex Trade-Offs
Multimodal AI: A Powerful Leap With Complex Trade-Offs

Forbes

time36 minutes ago

  • Forbes

Multimodal AI: A Powerful Leap With Complex Trade-Offs

Artificial intelligence is evolving into a new phase that more closely resembles human perception and interaction with the world. Multimodal AI enables systems to process and generate information across various formats such as text, images, audio, and video. This advancement promises to revolutionize how businesses operate, innovate, and compete. Unlike earlier AI models, which were limited to a single data type, multimodal models are designed to integrate multiple streams of information, much like humans do. We rarely make decisions based on a single input; we listen, read, observe, and intuit. Now, machines are beginning to emulate this process. Many experts advocate for training models in a multimodal manner rather than focusing on individual media types. This leap in capability offers strategic advantages, such as more intuitive customer interactions, smarter automation, and holistic decision-making. Multimodal has already become a necessity in many simple use cases today. One example of this is the ability to comprehend presentations which have images, text and more. However, responsibility will be critical, as multimodal AI raises new questions about data integration, bias, security, and the true cost of implementation. Multimodal AI allows businesses to unify previously isolated data sources. Imagine a customer support platform that simultaneously processes a transcript, a screenshot, and a tone of voice to resolve an issue. Or consider a factory system that combines visual feeds, sensor data, and technician logs to predict equipment failures before they occur. These are not just efficiency gains; they represent new modes of value creation. In sectors like healthcare, logistics, and retail, multimodal systems can enable more accurate diagnoses, better inventory forecasting, and deeply personalized experiences. In addition, and perhaps more importantly, the ability of AI to engage with us in a multimodal way is the future. Talking to an LLM is easier than writing and then reading through responses. Imagine systems that can engage with us leveraging a combination of voice, videos, and infographics to explain concepts. This will fundamentally change how we engage with the digital ecosystem today and perhaps a big reason why many are starting to think that the AI of tomorrow will need something different than just laptops and screens. This is why leading tech firms like Google, Meta, Apple, and Microsoft are heavily investing in building native multimodal models rather than piecing together unimodal components. Despite its potential, implementing multimodal AI is complex. One of the biggest challenges is data integration, which involves more than just technical plumbing. Organizations need to feed integrated data flows into models, which is not an easy task. Consider a large organization with a wealth of enterprise data: documents, meetings, images, chats, and code. Is this information connected in a way that enables multimodal reasoning? Or think about a manufacturing plant: how can visual inspections, temperature sensors, and work orders be meaningfully fused in real time? That's not to mention the computing power multimodal AI require, which Sam Altman referenced in a viral tweet earlier this year. But success requires more than engineering; it requires clarity about which data combinations unlock real business outcomes. Without this clarity, integration efforts risk becoming costly experiments with unclear returns on investment. Multimodal systems can also amplify biases inherent in each data type. Visual datasets, such as those used in computer vision, may not equally represent all demographic groups. For example, a dataset might contain more images of people from certain ethnicities, age groups, or genders, leading to a skewed representation. Asking a LLM to generate an image of a person drawing with their left hand remains challenging – leading hypothesis is that most pictures available to train are right-handed individuals. Language data, such as text from books, articles, social media, and other sources, is created by humans who are influenced by their own social and cultural backgrounds. As a result, the language used can reflect the biases, stereotypes, and norms prevalent in those societies. When these inputs interact, the effects can compound unpredictably. A system trained on images from a narrow population may behave differently when paired with demographic metadata intended to broaden its utility. The result could be a system that appears more intelligent but is actually more brittle or biased. Business leaders must evolve their auditing and governance of AI systems to account for cross-modal risks, not just isolated flaws in training data. Additionally, multimodal systems raise the stakes for data security and privacy. Combining more data types creates a more specific and personal profile. Text alone may reveal what someone said, audio adds how they said it, and visuals show who they are. Adding biometric or behavioral data creates a detailed, persistent fingerprint. This has significant implications for customer trust, regulatory exposure, and cybersecurity strategy. Multimodal systems must be designed for resilience and accountability from the ground up, not just performance. Multimodal AI is not just a technical innovation; it represents a strategic shift that aligns artificial intelligence more closely with human cognition and real business contexts. It offers powerful new capabilities but demands a higher standard of data integration, fairness, and security. For executives, the key question is not just, "Can we build this?" but "Should we, and how?" What use case justifies the complexity? What risks are compounded when data types converge? How will success be measured, not just in performance but in trust? The promise is real, but like any frontier, it demands responsible exploration.

Shiba Inu Price Sell-Off Continues as SHIB Burn Rate Skyrockets to 112,000%
Shiba Inu Price Sell-Off Continues as SHIB Burn Rate Skyrockets to 112,000%

Yahoo

timean hour ago

  • Yahoo

Shiba Inu Price Sell-Off Continues as SHIB Burn Rate Skyrockets to 112,000%

Shiba inu's (SHIB) supply-side dynamics are screaming bullish, yet the second-largest joke cryptocurrency by market value trades under pressure. Early this week, SHIB's burn rate surged to over 112,000%, with more than 116 million coins transferred to wallets that cannot spend money. In other words, these coins were permanently taken out of circulation. The daily burn rate refers to the number of SHIB tokens permanently destroyed or removed from circulation each day. Token burns are designed to decrease the supply of the cryptocurrency over time, bringing a deflationary appeal to the digital asset. "Over 527 trillion SHIB tokens are approaching profitability, while the burn rate exploded 112,839% with 116 million tokens removed from circulation," CoinDesk's AI insights noted. Furthermore, SHIB's ecosystem fundamentals demonstrated strength, with record wallet growth exceeding 1.5 million unique addresses and significant increases in Shibarium layer-2 transactions. Still, the memecoin remained locked in a downtrend at press time, last changing hands at $0.00001190, representing a 2% drop over the past 24 hours and a nearly 5% decline for the week. Overnight, the token faced strong selling pressure, with above-average volume exceeding 500 billion units, establishing resistance around $0.0000122. The double-bottom pattern is forming on charts, signalling a potential 20% rally to $0.000016. Key resistance has been established at $0.0000122, backed by above-average volumes. The narrow trading range ($0.00001203-$0.000012) indicates the consolidation phase. Volume spikes at 07:35 and 07:46-07:47 coincided with price recovery attempts. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data

The Browser Company launches its AI-first browser, Dia, in beta
The Browser Company launches its AI-first browser, Dia, in beta

Yahoo

timean hour ago

  • Yahoo

The Browser Company launches its AI-first browser, Dia, in beta

Traditional web tools are facing an existential crisis as AI products and tools increasingly eat up attention — and therefore market share and money — from a wide swathe of products that people have used for years to interact with the internet. At least, that's what The Browser Company seems to think is happening. The company last year decided to stop developing its popular web browser Arc, acknowledging that while Arc was popular among enthusiasts, it never hit scale as it presented too steep a learning curve to reach mass adoption. The startup has since been heads-down on developing a browser that bakes in AI at the heart of the browser. That browser, called Dia, is now available for use in beta, though you'll need an invite to try it out. The Browser Company's CEO Josh Miller has of late acknowledged how people have been using AI tools for all sorts of tasks, and Dia is a reflection of that. By giving users an AI interface within the browser itself, where a majority of work is done these days, the company is hoping to slide into the user flow and give people an easy way to use AI, cutting out the need to visit the sites for tools like ChatGPT, Perplexity and Claude. Up front, Dia presents a straightforward interface. The browser is based on Chromium, the open-source browser project backed by Google, so it has a familiar look and feel. The marquee feature here is the AI smarts, of course. Besides letting you type in website names and search terms, Dia's URL bar acts as the interface for its in-built AI chatbot. The bot can search the web for you, summarize files that you upload, and can automatically switch between chat and search functions. Users can also ask questions about all the tabs they have open, and the bot can even write up a draft based on the contents of those tabs. To set your preferences, all you have to do is talk to the chatbot to customize its tone of voice, style of writing, and settings for coding. Via an opt-in feature called History, you can allow the browser to use seven days of your browsing history as context to answer queries. Another feature called Skills lets you build small snippets of code that act as shortcuts to various settings. For example, you can ask the browser to build a layout for reading, and it'll code something up for you — think Siri shortcuts, but for your browser. Now, we have to note that chatbots in browsers are not a new feature at all. Several browser companies have integrated AI tools into their interfaces — for example, Opera Neon lets users use an AI agent to build mini-applications or complete tasks on their behalf, and Google is also adding AI-powered features to Chrome. The Browser Company says all existing Arc members will get access to Dia immediately, and existing Dia users will be able to send invites to other users. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into the world of global news and events? Download our app today from your preferred app store and start exploring.
app-storeplay-store