#60 - Beyond the Tipping Point
Image Credit: https://unsplash.com/@bernardhermant

#60 - Beyond the Tipping Point

Even though some changes have been talked about for years, we often don't realize it when they happen. For the first time in the U.S., linear TV and audio have both fallen below 50%, as per Nielsen and Edison Research. Linear TV's viewership was 63.6% in June 2021 and has now dropped to 49.6%. Similarly, linear audio went from 62% in 2019 to 49.7% in 2023. In the blink of an eye, the media landscape has fundamentally changed.

Similar numbers across Europe are hard to find, but in 2022 a Samsung report noted that streaming surpassed linear viewing on their devices in Europe for the first time. During H1 2022, linear made up 48% of total watch hours, while streaming constituted 52%.

This isn't a zero-sum game. It's not merely about on-demand platforms expanding endlessly while linear formats disappear. Both formats have their unique value propositions and are likely to coexist for a long time, serving different audience segments.

However, to remain relevant, media organizations need to adapt to the reality beyond this tipping point.

IN FOCUS

The EU's Digital Services Act (DSA) has officially come into effect this week, bringing with it significant changes for major platforms like TikTok, Instagram, and YouTube. These platforms now have to navigate stricter regulations, all designed to enhance user safety, especially for younger audiences.

Several platforms have made updates to comply with the terms of the DSA.

AI/NEWS

A US judge has ruled that AI-generated art isn't copyrightable, emphasizing the need for human involvement. This doesn't mean that an AI-illustration will be denied copyright just because it’s made with a machine, it just means that only the human involved can register a copyright, not the machine.

Related: Google and YouTube are trying to have it both ways with AI and copyright. Read more.

AI unlocks new ways to hack computers. A team of researchers from British universities has trained a deep learning model that can steal data from keyboard keystrokes recorded using a microphone with an accuracy of 95%.


AI/TOOLS & RESOURCES

  • What does artificial intelligence (AI) look like? Google's Deepmind has commissioned artists from around the world to create more diverse and accessible representations of AI. Link.
  • ElevenLabs, the AI-powered platform for creating synthetic voices launched its platform out of beta. Link.
  • Meta launched a new multimodal language model, SeamlessM4T, that translates 100 languages any way you want. Link.
  • Meta also released Code Llama, a new model optimized for generating code. Link.
  • OpenAI is letting users finetune ChatGPT-3.5 with their own data so it can execute more specific use cases. Link.
  • Twilio is providing an online tool to generate "nutrition labels" for AI, clearly outlining how data will be used. Link.
  • Midjourney, the top AI art platform, now offers inpainting, which lets users modify a portion of an image they wish to change. Link.
  • And Gartner released its latest Hype Cycle placing Generative AI at the peak of inflated expectations. Link.

NEWS

Researchers in Munich and London studied what readers want from data journalism. The report’s findings suggest audiences want data journalism to be concise, provide analysis, have a human angle and include visual elements and dislike too many numbers and complicated technical terms.

Another new study finds that people mainly share misinformation on social media because of the reward structures that reward attention rather than accuracy. The good news? Changing the incentives could change the behavior.

+ A new analysis on how malicious sites use AI chatbots to rewrite news articles that first appeared in outlets like CNN, The New York Times, and Reuters, without credit.

+ How to take the BBC’s 50:50 The Equality Project beyond improving representation in terms of numbers, to challenging stereotypical portrayal of women in news.

VIDEO

TikTok is revolutionizing film promotion, as showcased by the "Barbie" movie. Before hitting theaters, its teaser on TikTok garnered 27 million views and 2.7 million likes. By summer, the film's TikTok account had 1.4 million followers. Read more.

The other go-to-platform for teenagers, YouTube, is rapidly overshadowing traditional TV, as revealed by a Precise TV report. A staggering 85% of 13- to 15-year-olds in the US recently watched YouTube, while only 43% viewed broadcast TV. Read more.

GAMES

Reporters Without Borders created a library in Minecraft called "The Uncensored Library". Its purpose is to offer access to banned texts in nations where governments control social platforms.

Netflix’s newest gaming experiment, Netflix Stories, is allowing subscribers to become playable characters in their favorite shows.

The Metropolitan Museum of Art and Roblox have created Replica, an AR app that's linked to Roblox and allows users to dress their avatars with items from paintings throughout the museum, e.g., Van Gogh's straw hat.

And filmmaker Harmony Korine is taking his experimental, narrative-eschewing style to the next level with his new venture, Edglrd — a collective of artists creating hyperactive, interactive, and gamified entertainment.

SHORT


IDEAS, TRENDS & SHIFTS

Google Arts & Culture turned to Leonardo Da Vinci, featuring thousands of his sketches, 3D-versions of his inventions, macro-shots of his art and much more.

Frieze has a good long-read on the rise of generative AI and how the role of the artist might change if individuality and authenticity are understood to be faked.

New tech media outfit 404 lifts the lid off the dark underside of the generative AI web. Inside the AI porn marketplace where everything and everyone is for sale: (Warning: explicit and disturbing content - I doubted whether this belongs here, but it is important to also point out how AI tools that are promoted 'to do anything' will also push boundaries that we do not want to cross. Link.)


That's all for this week. Feel free to share with friends and colleagues!

Ezra

?







要查看或添加评论,请登录

Ezra Eeman的更多文章

  • #119 - Assistants over Articles

    #119 - Assistants over Articles

    I was invited this week to an Aspen Digital meeting in London called AI & News: Charting the Course. It was a Chatham…

    1 条评论
  • #118 - Browsers with Brains

    #118 - Browsers with Brains

    The web browser, that familiar rectangular interface that has remained fundamentally unchanged for decades, stands at…

    5 条评论
  • #117 - All-in-one Models

    #117 - All-in-one Models

    The AI landscape shifts weekly — a constant parade of new models — yet the actual progress is still incremental…

    3 条评论
  • #116 - An Inconvenient Truth

    #116 - An Inconvenient Truth

    AI chatbots are often promoted as the ultimate answer machines, and with billions invested, admitting their flaws…

    1 条评论
  • #115 - Lost in Model Land

    #115 - Lost in Model Land

    The AI landscape has transformed into a maze of specialized models, each more sophisticated than its predecessor. We…

    1 条评论
  • #114 - Community in the Loop

    #114 - Community in the Loop

    When we discuss AI in newsrooms, the conversation often gets stuck in one dimension: journalists versus machines…

    1 条评论
  • #113 - Made for Machines

    #113 - Made for Machines

    Imagine teaching a robot to post on Instagram. Would you really build one with eyes and hands, train it to tap screens…

    3 条评论
  • #112 - Platformageddon

    #112 - Platformageddon

    The obvious headline this week is TikTok's dramatic shutdown in the U.S.

    4 条评论
  • #111 - Great Expectations

    #111 - Great Expectations

    Wayfinder is back after a few weeks of total rest and reading – and 2025 has already been shaped by Mark Zuckerberg's…

  • #110 - 9 Dynamics Reshaping Media

    #110 - 9 Dynamics Reshaping Media

    This blog post is an AI-generated summary of my recent Wayfinder Report. The summary was created using NotebookLM, and…

    5 条评论

社区洞察

其他会员也浏览了