#106 - The Value Question

When it comes to artificial intelligence, we often get caught up in the wrong conversation. I am not entirely free of guilt, focusing perhaps too much on new AI models and their growing capabilities – from early 'reasoning' to real-time multimodal output. But the real story is, of course, not about what AI can do; it's about where people find genuine value in using it.

Recent studies* show that even among teens, who we might expect to be early adopters, only a quarter regularly use AI tools like ChatGPT, Gemini or Copilot. Tellingly, when they do use it, it's primarily for homework help – where the benefit is clear and immediate. Fewer youth turn to generative AI for personal activities, suggesting these tools haven't yet demonstrated compelling value in their daily lives beyond assigned tasks.

This gap between capability and purpose-driven adoption echoes what Rhea Purohit explores in her Learning Curve newsletter this week: the success of any technology isn't just about its technical capabilities. Instead, it depends on four key human factors: expected benefits, ease of use, social influence, and available support. In other words: AI's impact will not be shaped by its capabilities alone, but by how clearly it demonstrates value in both our work and personal lives.

*Both the News Literacy Project and Common Sense Media report similar findings: roughly a quarter of American teens (23-24%) engage with AI chatbots on a weekly basis.

In Focus

Donald Trump's return to the US presidency has set off a flood of debates, opinion pieces, and analysis over the past few days. The Reuters Institute has put together a good overview, with views ranging from 'traditional journalism is dead' to 'we're heading for trouble now that Trump is back'.

According to some observers, Donald Trump's early and heavy investment in relationships with podcasters and livestreamers proved to be a winning strategy. However, critics like Max Read are more skeptical, warning against jumping to easy conclusions about a supposed 'TikTok electorate'.

AI in News

AI companies like OpenAI, Google, and Meta are more heavily reliant on premium publisher content for training their LLMs than publicly acknowledged, according to new research. However, so far evidence has not convinced judges. This week two news outlets lost their copyright lawsuit against OpenAI.

UK's Reach has integrated AI into 25% of its content production, with their internal AI-bot Guten dramatically reducing breaking news publishing time from 9 minutes to just 90 seconds. This makes Reach one of the first news organizations to demonstrate measurable benefits from generative AI at scale.

Short:

• LLMs help New York Times reporters transcribe and sort through hundreds of hours of leaked audio. More
• Brazilian fact-checking powerhous Aos Fators launches free AI-powered fact-checking services. Check it here
• The Washington Post debuted “Ask The Post AI” a generative AI tool leveraging the publication’s journalism to deliver summary answers and curated results. More info + Try it here
• Google reports its AI Overviews in search now reaches 1 billion monthly users. Link
• Medium faces challenges with AI-generated content flooding the platform. Read more
• Many publishers are increasingly reliant on Google Discover for their daily traffic numbers accroding to PressGazette. It’s a risky bet.

Other AI

The term “agents” is somewhat of a buzzword that doesn’t really acknowledge the many different types of agents.?This guide lays out 10 kinds of generic agents, what they can do, how much human interaction is required and how feasible they are.

Short:

• Mistral launched a new Content Moderation tool to detect harmful content in text across multiple languages. Link
• Good piece on MIT's Tech Review: "AI search could deprive creators and publishers of the visits and eyeballs they need to survive". Read
• Ethan Mollick considers the potential impact of AI’s current capabilities like reasoning and real-time generation. More
• INMA's Jodie Hopperton on how the rollout of new AI technologies is changing the user experience. Deep dive

Quick Hits

• Disney unveils a new business unit to explore opportunities and risks for AI and XR.
• Video model Kling AI 1.5 lets you generate videos with consistent-looking characters.
• Amazon’s Prime Video launches AI-generated TV Show recaps.
• Spotify has rolled out an easier way to share podcasts to TikTok.
• The U.K. Government proposes changes to media merger laws for the digital age.

TRENDS

Doug Shapiro sees four major forces reshaping media: fragmentation, disintermediation, concentration (of power and attention), and virtualization. These tectonic shifts are redistributing value among creators, traditional and new intermediaries, and consumers.

In a new report BCG values the personalization market at $2 trillion. Their research suggests companies can capture their share of this opportunity by following specific AI-driven strategies for customer engagement.

And less than 2 in 10 teens in the US can distinguish between different types of information, such as news, advertisement, opinion, entertainment according to a new study by the News Literacy Project.

READ

Deloitte has mapped out four possible futures for generative AI, based on two key variables: AI performance and the distribution of economic benefits (shared vs. unequal growth).

AI isn't just algorithms and code - its physical footprint spans the globe, from mines in Nevada and the Democratic Republic of Congo to data centers in Virginia. At every stage, its massive appetite for energy and water is reshaping communities worldwide.

Long read of the week: A speculative vision of how AI and machine-learning systems could impact and influence art and creative work for the next 10 years.

+ The Guardian calculates our daily digital activities' carbon costs, from email to messaging.

?That's it for this week - as always, Wayfinder is made to travel so feel free to share it with friends and colleagues!