"Mr Branding" is a blog based on RSS for everything related to website branding and website design, it collects its posts from many sites in order to facilitate the updating to the latest technology.
To suggest any source, please contact me: Taha.baba@consultant.com
The US Federal Trade Commission has opened an inquiry into how major technology companies manage the risks of AI chatbots, focusing on their impact on younger users.
Alphabet, Meta, OpenAI, Snap, X, Instagram, and Character Technologies have been ordered to provide detailed information about their chatbot operations. The request covers product design, safety testing, data handling, monetization practices, and the way chatbot characters are developed. The companies must respond by late September.
Why the Investigation Was Launched
The move follows reports of troubling exchanges between AI systems and children. Meta’s chatbots were accused of allowing inappropriate conversations with minors. Snapchat’s “My AI” drew criticism for its interactions with younger users. X’s recently launched chatbot companions also raised concerns about how people may form personal attachments to these digital agents.
The FTC is seeking clarity on how companies measure and limit risks, particularly when chatbots act as companions. It wants to know whether safeguards are built into these products, whether companies restrict use by minors, and how users and parents are informed about potential dangers. The order also asks for details on how inputs and outputs are processed and how safety evaluations are conducted.
Although the inquiry is not tied to a specific enforcement action, the Commission has signaled that the information will guide future decisions on consumer protection and child safety.
Balancing Regulation and Innovation
Officials have said that protecting children online is a priority while also noting the need for the United States to maintain leadership in AI development. The outcome of the study may shape how these goals are balanced.
The investigation recalls earlier debates over social media oversight, when warnings about youth safety emerged long before strong rules were introduced. The findings of this review could influence how chatbot providers are required to manage their systems in the years ahead.
Notes: This post was edited/created using GenAI tools. Image: DIW-Aigen.
Artificial intelligence is a common feature in new smartphones, but most buyers continue to focus on practical needs. A survey of more than 2,200 adults in the United States found that only 11 percent of smartphone owners would upgrade their device because of AI features. That figure has fallen seven points from last year. Around three in ten people also said AI on phones is not useful and they would prefer fewer additions.
Companies Push AI, Buyers Look Elsewhere
The findings come as major manufacturers continue to expand AI across their product lines. Samsung included Galaxy AI in the S25 series and its latest foldable devices. Google built Gemini into the Pixel 10 range. Apple introduced Apple Intelligence on recent models and is expected to reveal further updates during the iPhone 17 launch. Despite these investments, the survey shows that AI is not driving most purchase decisions.
What Buyers Want in a Phone
When asked what matters most, 62 percent named price as the key factor. Battery life followed at 54 percent, and extra storage at 39 percent. Camera performance remains another concern, mentioned by three in ten respondents. The same areas topped the list in last year’s survey, which showed battery life, storage, and camera as the strongest reasons for an upgrade.
Thin Designs and Other Trends
Industry events this year placed emphasis on slimmer phones. Samsung presented the S25 Edge, while Oppo and Honor showed thinner foldable concepts. Only seven percent of survey participants said a thinner design would encourage them to upgrade. The results suggest the appeal of new form factors is limited outside trade shows and product launches.
Limited Use of AI Tools
Although most smartphones now include AI, few owners use it regularly. Only 13 percent said they rely on AI for text summaries or writing. Eight percent use image generation, and seven percent use AI for photo editing. Twenty percent admitted they are unsure how to use the AI tools built into their phones.
Privacy and Cost Concerns
Privacy is a growing worry. Just over 40 percent of respondents said they are concerned about how AI handles their data, up seven points from the previous year. Cost is another barrier. Half of the sample said they would not pay extra for AI features, even as companies consider charging for access.
Interest Varies by Age
Younger groups showed greater interest in mobile AI. One in four Gen Z respondents said they find the features helpful. Sixteen percent of millennials agreed. Among iPhone owners, 61 percent use Siri, though only one in ten rely on it daily. On Google’s Pixel devices, 41 percent use Gemini, but daily use is again limited.
Outlook
The survey results point to a clear trend. AI has become a standard feature across premium and mid-range devices, yet buyers continue to prioritize price, battery performance, storage, and camera quality when deciding whether to upgrade. For now, companies face the challenge of justifying their heavy investment in AI while most customers remain focused on the basics.
For creators hoping to reach audiences beyond their home market, YouTube is rolling out fresh options that make content more accessible in multiple languages. The platform is extending its multi-language audio system to all channels, along with early tests of localized video thumbnails.
The audio feature lets creators attach additional tracks in different languages to a single upload. That flexibility gives them control over how translations are delivered, whether through their own recordings, professional voice artists, or artificial intelligence services. By contrast, YouTube’s automated dubbing tool, which is still expanding separately, relies entirely on AI-generated speech.
Results from the pilot program, which began in 2023, suggest clear benefits. Some participants saw a significant portion of their total watch time coming from viewers outside the main language of the channel. Entertainment and food-focused videos performed especially strongly once dubbed versions were added.
Importantly, translated tracks are treated as part of the same upload rather than competing with the original version. That approach allows creators to revisit their best-performing clips, add alternative audio, and expand reach without losing visibility in recommendations.
In addition to audio, YouTube has been experimenting with localized thumbnails. Selected partners have been testing preview images that adjust text according to a viewer’s preferred language. The idea is to make videos feel tailored for different markets and to improve first impressions for international viewers.
Taken together, these changes show YouTube’s effort to strengthen its role as a global platform, giving creators more tools to connect with audiences far beyond their initial community.
Artificial intelligence is no longer a side note in social media. In 2025, it has become the clearest dividing line between trusted brands and those audiences question. New findings from SproutSocial’s Q2 and Q3 Pulse Surveys, spanning more than 2,000 respondents in the US, UK, and Australia, underscore just how much AI shapes user expectations.
AI-generated content disclosure leads all concerns
When asked what worries them most about brand behavior on social platforms, respondents put undisclosed AI-generated content at the top of the list, ahead of data mishandling.
Gen Z and Millennials are the most likely to rank AI disclosure as their number-one issue.
Older generations place stronger emphasis on data privacy, but disclosure still features prominently across groups.
Human-made content holds the trust advantage
Audiences show a measurable bias toward human creators:
55% of consumers say they are more likely to trust brands that commit to publishing content created by people rather than AI.
This rises to 62% among Millennials.
Among those who identify as Liberal, 61% prefer brands that publish human-generated content.
Respondents also value attribution. Brands that credit photographers, stylists, and makers behind their posts gain credibility, reinforcing the idea that transparency is as important as output.
AI influencers remain divisive
Synthetic influencers highlight the sharpest tension between innovation and trust:
46% of consumers say they are not comfortable with brands using AI-generated influencers.
31% say their comfort depends on the campaign.
By demographics: 28% of men say they are comfortable, versus 19% of women.; 32% of Gen Z and 33% of Millennials express comfort, compared with only 23% of Gen X.
These splits show some openness among younger audiences, but discomfort remains the dominant stance.
What the numbers mean for brands
The data suggests that AI is acceptable to audiences only under clear conditions: it must be disclosed, human creators must be acknowledged, and synthetic influencers remain a risk for brand perception. Consumers are not rejecting AI entirely; they are demanding honesty and choice.
The wider landscape in brief
Beyond AI, the same surveys report that 60% of users say social media improved their mental health in the past six months, and 76% say social influenced a purchase during that period. Younger generations also continue to see platforms as tools for financial education. But none of these figures resonate as strongly with trust as the numbers tied to AI.
Notes: This post was edited/created using GenAI tools.
August 2025 App Rankings Show Slower Downloads and Steady Revenues
Mobile app rankings for August reveal a quieter month. According to Appfiguresdata, downloads fell compared with July, while revenues stayed firm. Market leaders held their ground, and streaming platforms gained more space in the earnings chart. The results underline seasonal shifts as users returned from vacations and spending patterns settled into late summer routines.
Downloads Slip as Vacations End
ChatGPT led global downloads in August with about 48 million installs across iOS and Android. Most of its growth came from Google Play, where it added 35 million downloads. Apple’s App Store contributed 13 million. TikTok followed with 38 million installs, supported by its Lite version on Android. Instagram remained close behind with 35 million. Facebook and WhatsApp completed the global top five with 27 million and 24 million installs.
The combined figure for the ten most downloaded apps reached 273 million in August. This marked a decline from July, reflecting reduced activity during the vacation period. Threads moved higher in the overall chart, and Uber entered the iOS top ten for the first time this year.
iOS Download Chart
On the App Store, ChatGPT led with 13 million installs. Threads ranked second with 11 million. Google Maps, CapCut, and Temu followed, each recording about 8 million downloads. Google’s search app and TikTok tied with 7 million. Telegram added 6 million. Uber and Google Chrome closed the list with 5 million each.
Google Play Download Chart
On Google Play, ChatGPT topped the list with 35 million installs. TikTok Lite came second with 32 million, while Instagram reached 30 million. Facebook and WhatsApp followed with 23 million and 19 million. Temu and Snapchat both added 15 million downloads. WhatsApp Business and Telegram each brought in 13 million. CapCut ranked tenth with 12 million installs.
Combined Global Downloads
When both platforms are combined, ChatGPT held first with 48 million installs. TikTok reached 38 million, Instagram 35 million, Facebook 27 million, and WhatsApp 24 million. Temu stood at 22 million, Threads at 21 million, and CapCut at 20 million. Telegram followed with 19 million, while Snapchat closed the top ten with 18 million.
Rank
iOS App Store (Downloads)
Google Play
Combined Total
1
ChatGPT – 13M
ChatGPT – 35M
ChatGPT – 48M
2
Threads – 11M
TikTok Lite – 32M
TikTok – 38M
3
Google Maps – 8M
Instagram – 30M
Instagram – 35M
4
CapCut – 8M
Facebook – 23M
Facebook – 27M
5
Temu – 8M
WhatsApp – 19M
WhatsApp – 24M
6
Google – 7M
Temu – 15M
Temu – 22M
7
TikTok – 7M
Snapchat – 15M
Threads – 21M
8
Telegram – 6M
WhatsApp Business – 13M
CapCut – 20M
9
Uber – 5M
Telegram – 13M
Telegram – 19M
10
Google Chrome – 5M
CapCut – 12M
Snapchat – 18M
Revenue Rankings Hold Steady
While downloads slowed, earnings remained stable. TikTok was again the top revenue generator in August. The app brought in an estimated $335 million after store fees, keeping a clear lead over other services. ChatGPT followed with $170 million, supported mainly by App Store purchases. YouTube held third place with $137 million, relying on iOS transactions rather than standard billing on Android.
Tinder and Disney+ secured fourth and fifth with $130 million and $115 million. HBO Max and Google One posted about $100 million each. CapCut earned $93 million. Peacock TV and Crunchyroll completed the chart with $59 million and $58 million. Together, the ten highest-earning apps generated about $1.3 billion, a slight increase from July.
iOS Revenue Chart
TikTok topped iOS with $213 million in net revenue. YouTube followed with $137 million, and ChatGPT ranked third with $123 million. Tinder added $92 million, while CapCut produced $81 million. Disney+ earned $76 million and HBO Max $72 million. Snapchat contributed $49 million. Peacock TV and Tencent Video each recorded $48 million.
Google Play Revenue Chart
On Google Play, TikTok earned $123 million, the highest on the platform. Google One followed with $99 million. ChatGPT generated $47 million. Amazon’s mobile app was close behind with $46 million. Disney+ collected $40 million, while Tinder reached $37 million. Spotify earned $32 million and HBO Max $29 million. Crunchyroll and Prime Video shared the final places, both at $20 million.
Combined Global Revenue
The combined rankings placed TikTok far ahead with $335 million. ChatGPT held second at $170 million. YouTube stood at $137 million. Tinder and Disney+ followed with $130 million and $115 million. HBO Max reached $101 million, while Google One added $100 million. CapCut ranked eighth with $93 million. Peacock TV recorded $59 million and Crunchyroll $58 million.
Rank
iOS App Store (Revenue)
Google Play
Combined Total
1
TikTok – $213M
TikTok – $123M
TikTok – $335M
2
YouTube – $137M
Google One – $99M
ChatGPT – $170M
3
ChatGPT – $123M
ChatGPT – $47M
YouTube – $137M
4
Tinder – $92M
Amazon – $46M
Tinder – $130M
5
CapCut – $81M
Disney+ – $40M
Disney+ – $115M
6
Disney+ – $76M
Tinder – $37M
HBO Max – $101M
7
HBO Max – $72M
Spotify – $32M
Google One – $100M
8
Snapchat – $49M
HBO Max – $29M
CapCut – $93M
9
Peacock TV – $48M
Crunchyroll – $20M
Peacock TV – $59M
10
Tencent Video – $48M
Prime Video – $20M
Crunchyroll – $58M
Market Trends
Seasonality shaped August’s results. Downloads slowed, but revenue stayed stable, with most mid-tier apps earning more than in July. Streaming platforms performed strongly, with four services in the revenue chart compared with the usual two. These included Disney+, HBO Max, Peacock TV, and Crunchyroll, all supported by summer demand for entertainment.
Outlook
As September begins, the market will show whether download activity returns to July’s levels. The revenue side suggests that subscription models and media services continue to hold ground. TikTok, ChatGPT, and YouTube remained the strongest performers across both categories, reinforcing their positions as leading players in the global app economy.
Notes: This post was edited/created using GenAI tools.
Attempts at communicating what generative artificial intelligence (AI) is and what it does have produced a range of metaphors and analogies.
From a “black box” to “autocomplete on steroids”, a “parrot”, and even a pair of “sneakers”, the goal is to make the understanding of a complex piece of technology accessible by grounding it in everyday experiences – even if the resulting comparison is often oversimplified or misleading.
Image: unsplash
One increasingly widespread analogy describes generative AI as a “calculator for words”. Popularised in part by the chief executive of OpenAI, Sam Altman, the calculator comparison suggests that much like the familiar plastic objects we used to crunch numbers in maths class, the purpose of generative AI tools is to help us crunch large amounts of linguistic data.
The calculator analogy has been rightlycriticised, because it can obscure the more troubling aspects of generative AI. Unlike chatbots, calculators don’t have built-in biases, they don’t make mistakes, and they don’t pose fundamental ethical dilemmas.
Yet there is also danger in dismissing this analogy altogether, given that at its core, generative AI tools are word calculators.
What matters, however, is not the object itself, but the practice of calculating. And calculations in generative AI tools are designed to mimic those that underpin everyday human language use.
Languages have hidden statistics
Most language users are only indirectly aware of the extent to which their interactions are the product of statistical calculations.
Think, for example, about the discomfort of hearing someone say “pepper and salt” rather than “salt and pepper”. Or the odd look you would get if you ordered “powerful tea” rather than “strong tea” at a cafe.
The rules that govern the way we select and order words, and many other sequences in language, come from the frequency of our social encounters with them. The more often you hear something said a certain way, the less viable any alternative will sound. Or rather, the less plausible any other calculated sequence will seem.
In linguistics, the vast field dedicated to the study of language, these sequences are known as “collocations”. They’re just one of many phenomena that show how humans calculate multiword patterns based on whether they “feel right” – whether they sound appropriate, natural and human.
Why chatbot output ‘feels right’
One of the central achievements of large language models (LLMs) – and therefore chatbots – is that they have managed to formalise this “feel right” factor in ways that now successfully deceive human intuition.
In fact, they are some of the most powerful collocation systems in the world.
By calculating statistical dependencies between tokens (be they words, symbols, or dots of color) inside an abstract space that maps their meanings and relations, AI produces sequences that at this point not only pass as human in the Turing test, but perhaps more unsettlingly, can get users to fall in love with them.
A major reason why these developments are possible has to do with the linguistic roots of generative AI, which are often buried in the narrative of the technology’s development. But AI tools are as much a product of computer science as they are of different branches of linguistics.
The ancestors of contemporary LLMs such as GPT-5 and Gemini are the Cold War-era machine translation tools, designed to translate Russian into English. With the development of linguistics under figures such as Noam Chomsky, however, the goal of such machines moved from simple translation to decoding the principles of natural (that is, human) language processing.
The process of LLM development happened in stages, starting from attempts to mechanise the “rules” (such as grammar) of languages, through statistical approaches that measured frequencies of word sequences based on limited data sets, and to current models that use neural networks to generate fluid language.
However, the underlying practice of calculating probabilities has remained the same. Although scale and form have immeasurably changed, contemporary AI tools are still statistical systems of pattern recognition.
They are designed to calculate how we “language” about phenomena such as knowledge, behaviour or emotions, without direct access to any of these. If you prompt a chatbot such as ChatGPT to “reveal” this fact, it will readily oblige.
ChatGPT-5 response when asked if it uses statistical calculations to form its responses.OpenAI/ChatGPT/The Conversation
AI is always just calculating
So why don’t we readily recognise this?
One major reason has to do with the way companies describe and name the practices of generative AI tools. Instead of “calculating”, generative AI tools are “thinking”, “reasoning”, “searching” or even “dreaming”.
The implication is that in cracking the equation for how humans use language patterns, generative AI has gained access to the values we transmit via language.
But at least for now, it has not.
It can calculate that “I” and “you” is most likely to collocate with “love”, but it is neither an “I” (it’s not a person), nor does it understand “love”, nor for that matter, you – the user writing the prompts.
Generative AI is always just calculating. And we should not mistake it for more.
Amnesty International has documented how Pakistan has built a wide-reaching surveillance and censorship system using technology supplied by companies across several regions. The investigation, carried out with media and research partners including Paper Trail Media, DER STANDARD, Follow the Money, The Globe and Mail, Justice for Myanmar, InterSecLab, and the Tor Project, found that tools from North America, Europe, China, and the Middle East form the backbone of Pakistan’s monitoring network.
At the core are two platforms, the Web Monitoring System (WMS 2.0) and the Lawful Intercept Management System (LIMS). Both were developed and upgraded through international suppliers. The first firewall arrived in 2018 with equipment from Canadian firm Sandvine, later renamed AppLogic Networks. Trade records show shipments to local contractors as early as 2017, including Inbox Technologies, SN Skies Pvt Ltd, and A Hamson Inc.
By 2023, WMS 2.0 replaced the original setup. This version was supplied by China’s Geedge Networks with additional components from Niagara Networks in the United States and Thales in France. Amnesty describes it as a commercial export of China’s Great Firewall.
LIMS was built with technology from German company Utimaco, distributed through UAE-based Datafusion. Together, these systems gave Pakistani authorities broad capacity to intercept and review traffic moving across national telecommunications networks.
Surveillance Reach
WMS 2.0 works as both a monitoring and blocking tool. Investigators found it could log emails, VoIP calls, and browsing sessions in real time. It also allows operators to block websites and virtual private networks. With deep packet inspection, data sent over the internet can be broken down and examined in detail. While HTTPS encryption masks content, operators can still identify websites visited and gather metadata.
The firewall runs across all major telecom providers and at Pakistan’s international internet landing station. This positioning ensures that most digital traffic entering or leaving the country is subject to monitoring.
LIMS extends surveillance to phone calls, text messages, and location data. Investigators found that state agencies can initiate monitoring with only a phone number. This provides access to communication logs and, when connections are unencrypted, entire pages of content. Even when encrypted, the system still records websites visited and times of access.
According to Amnesty, LIMS has the capacity to monitor more than four million users at once. Telecom operators are legally required to connect their systems as a condition of their license.
Public Financing
Expansion of the network has relied heavily on public funds. In November 2023, five billion rupees from the Universal Service Fund were redirected to the firewall project through the ICT R&D Fund, also known as Ignite. In February 2024, the Economic Coordination Committee approved another ten billion rupees under the Digital Information Infrastructure initiative, which includes WMS 2.0.
These allocations covered upgrades to the firewall and wider integration of LIMS. Amnesty’s report traces the procurement of equipment through trade databases, showing how suppliers across multiple jurisdictions provided hardware and software despite the human rights risks.
Oversight and Legal Gaps
Pakistan’s Fair Trial Act requires warrants for surveillance, but Amnesty found these safeguards are often bypassed. The Pakistan Telecommunication Authority has acknowledged in court that large-scale interception takes place.
Authorities frame the systems as lawful interception tools for national security. Critics argue that little transparency or independent oversight exists. Citizens are not notified when their data is monitored, and Amnesty concludes that in practice LIMS enables indiscriminate surveillance.
Effect on Citizens
Journalists and civil society groups report that the systems are already narrowing space for free expression. One journalist told investigators that after publishing on corruption, both he and his contacts faced greater scrutiny. The awareness of monitoring led him to self-censor and avoid open communication, even with family members.
Amnesty’s Secretary General Agnès Callamard compared the technologies to watchtowers that constantly observe citizens, warning that unchecked surveillance undermines civic space and democratic participation.
International Role
The investigation highlights the responsibility of companies and governments exporting surveillance tools. Of twenty companies contacted, only Niagara Networks in the United States and AppLogic Networks in Canada provided full responses. Utimaco and Datafusion initially answered in 2024 but did not follow up.
Out of nine government bodies approached, only Germany’s export authority and Canada’s trade bureau acknowledged correspondence. Neither gave substantive replies. Pakistan’s government did not respond.
Amnesty argues that this silence reflects a broader failure to regulate the trade of surveillance equipment. The report notes that suppliers in Europe, North America, China, and the UAE continue to export such tools without safeguards, even where they are used to restrict rights.
Expanding Framework
Together, the firewall and intercept platforms form a nationwide monitoring system. They rely on international technology, domestic funding, and limited oversight. Amnesty concludes that this framework leaves private communication exposed and narrows space for free expression.
Pakistan’s case is presented as part of a larger global debate on how surveillance technologies are exported and used. With procurement continuing and safeguards absent, the country’s monitoring capacity is expected to expand further, raising deeper concerns over privacy and basic freedoms.
