Over 100M Snapchatters Rely on AI for Conversation Openers, Claims Snapchat
Snapchat’s new ‘My AI’ feature has certainly caused a stir on social media, with many parents nationwide warning their children to stay well away from the feature amid reports of the AI producing alarming replies to younger audience queries. However despite its flaws, the feature has been incredibly popular, with more than 150 million people having sent over 10 billion messages to My AI, therefore Snapchat has decided to release a report delving into the 10 billion messages sent to the feature.
The standout statistic from the report was that over 100m messages were sent to ‘My AI’ asking for the best conversation starters, which was the most popular topic of conversation amongst the statistics. There have been over 25m conversations with the AI about dogs, alongside 46m conversations being based around football.
Interestingly, the AI also spurted out conversations about individual celebrities, having over 5m conversations about footballing icon Christiano Ronaldo, whilst also having 3m conversations with users about musicians such as Harry Styles, Billie Eilish and Lana Del Rey.
Find out more here.
Twitter To Suspend Accounts Found Guilty of ‘Trendjacking’
Once again making the headlines this week is boisterous billionaire and CTO of Twitter, Elon Musk. Recently, Musk has set his sights on combatting trendjacking, after responding to a user tweet asking about Musk’s thoughts on the matter, where Musk commented “Gaming replies to generate free advertising will result in account suspension”
When enforcing these new rules, Twitter needs to clarify exactly what trend jacking is, since, given Twitter’s nature, it’s not something that is seen as straightforward by its users. The official definition of meaning is ‘the practice of taking advantage of current events or news stories in such a way as to promote or advertise one’s product or brand’, however, it has long been a method to gain more exposure and promotion in the app, used by a lot of brands as a tactic to maximize awareness.
It is thought that Musk will target companies who are caught spamming promotions under unrelated tweets, which he later clarified in a tweet stating, “Accounts that try to game our verification system with non-sequitur self-promotion or advertise in a misleading way will be suspended”, essentially meaning that that replies which are out of context, and that seek to advertise a brand in a tweet’s replies, will now be suspended if caught.
Find out more here.
SquareSpace Acquires ‘Google Domains’ Amid Cost-Cutting Issues
Google has recently announced that it is selling its ‘Domains’ business to the website-building company, Squarespace for over £140 million, as a part of its cost-cutting efforts and focus on other areas of the digital world.
Google Domains, launched in 2014, offers custom URL identifiers with over 300 domain endings to enhance web presence. The service provides users with a domain which can be used to launch a website, create custom email accounts and integrate with third-party services. While Squarespace will support existing domains, the integration with Google’s sorting systems will be lost, potentially diminishing their value. It is also reported that Squarespace’s pricing may also be higher than Google Domains, which has unsettled a few current users of the service.
Earlier this year, Google announced a significant reduction in its workforce to redirect talent to emerging opportunities. Therefore this sale reflects Google’s ongoing change in strategy and its response to competition in generative AI and eCommerce.
Find out more here.
Meta Unveils New Speech-to-Text Tool ‘Voicebox’
Earlier this week, Meta announced the launch of its new ‘Voicebox’ AI system, which translates text into audio in a variety of styles and dialects. Voicebox is a system that translates text inputs into audio, enabling advanced text-to-audio translations with fewer learning and processing requirements than other similar systems.
In addition, Voicebox enables users to use any audio clip they like to make their text-to-speech translation sound like another person is speaking, by simply putting in a few seconds of audio. There is a concern among users of the platform that this could encourage a spike in deep fakes, with users experiencing similar incidents on TikTok with their existing text-to-speech feature, where celebrities’ voices were portrayed to quote things that they actually did not say.
Due to this worry, Meta isn’t releasing the source code or app to the public yet, citing ‘the potential risks of misuse’ when discussing the new feature. The company hopes instead to find more practical, valuable uses for the technology over time, and launch it later in the year.
Find out more here.