#88 - The Real Deal
Image: Lexica.art

#88 - The Real Deal

It's deal time. This week saw another round of partnerships between major publishers like the FT* and Axel Springer**, and AI companies. But make no mistake—these deals aren't (really) about news; they're about data.

Data, not news, is the ultimate prize for AI companies. It fuels AI models, and the quality, quantity, and diversity of data are crucial for building effective, unbiased systems. Hence, it's no surprise that AI companies are striking deals with key publishers as finding quality data becomes increasingly challenging. However, these deals aren't necessarily favorable to the whole news industry.

Ai-analyst Bret Kinsella thinks that major model developers are unlikely to continue licensing news content extensively once they've acquired enough high-quality training data. In fact, once they have a few deals in place with some key publishers and press agencies the value of other news content drops dramatically.

The deals are therefore at most a perk for some media companies, primarily serving the needs of AI companies. Similar to other tech partnerships, they may not last forever and they certainly aren't for everyone.


The Details:

*The deal between OpenAI and the FT is centered on enhancing ChatGPT with attributed content, improving the model and developing new AI products.

**The deal between Microsoft and Axel Springer is primarily an adtech partnership that also includes an AI news experiences component.

***Google is paying News Corp up to US$6 million a year in the US to develop new AI-related content and products.

AI

The marriage of social media and AI is not without its challenges. Early adopters find the integration of AI chats in Facebook and Instagram frustrating , making the apps difficult to use. While the bots are intriguing, they're unreliable. Additionally, a surge of AI-generated content from users is filling feeds, turning these platforms into a 'zombie internet'.

A new AI survey by the Verge suggests that 61% of Gen Z and 53% of Millennials use AI tools over traditional search engines like Google for research. Given the dramatic nature of these stats, it's hard not to question their accuracy.

Related: A deep dive into clickstream data sees no significant shift in search yet. Google continues to dominate the majority of search volume over all.

+ A new study by McAfee and Google suggests that worries about AI causing widespread job losses might be exaggerated. However, the study does not really substantiate this with new arguments or data.

Short updates:

  • Apple intensifies talks with OpenAI for generative AI features. Link
  • Curio is building an ‘AI news anchor’ in an app. Link
  • BBC is investing in AI to help transform its education services. Link
  • Anthropic launched an iOS app for Claude. Link
  • Microsoft releases its first responsible AI and transparency report. Link
  • OpenAI is rumored to launch an own search engine next week. Link

NEWS

Making money from generative AI chatbots isn't taking off as quickly as some hoped. Some publishing execs are holding off on adding ads to these chatbots. They prefer to wait until there's enough user engagement or they can develop a subscription model.

In a major new study , Schibsted has identified four key drivers of media trust: process credibility, content credibility, personal relevance, and selectivity.

Short:

  • David Caswell considers how Gen AI is redesigning news media companies. Link
  • Mattia Peretti discusses integrating AI in journalism without sacrificing integrity. Link
  • RISJ takes a look at how fact-checkers apply AI to their work. Link
  • Related: a practical guide for journalists and fact-checkers to identify machine-written content. Link

SOCIAL MEDIA

Researchers are finding that social media tech isn't just influencing cultural attitudes, it's actually shaping generations. Teens feel that social media algorithms, like those on TikTok, really get them.

The latest Digital 2024 Global Statshot report is out : Just one of the many insights in the report: The average duration of a single session of using YouTube is 7 mins 37 secs. For TikTok it's 5m 52 secs, and for Instagram it's 2 mins 51 secs.

And analyst Ben Thompson considers the long term future of Meta.

VIDEO

In 2023, more people created content on YouTube than ever before. YouTube Shorts, in particular, is really taking off. The number of channels uploading these short videos jumped by 50% from last year , and recently, the platform surpassed 3 million channels uploading Shorts.

Short

  • Amazon and YouTube are rumored to be frontrunners to snap up streaming rights to the NBA. Link
  • How to make a movie trailer using AI tools. Link
  • Dan Taylor compares AI Video generators. Link
  • Kocowa+, a Korean niche streamer specialised in K-content (Korean drama & tv-series) is launching in 39 countries across Europe and Oceania. Link

AUDIO

Gen Z is really getting into video podcasts, especially on YouTube. According to Edison Research, 16 out of the top 30 podcasts last year were available in video format, which is a big jump from just seven in 2021. And now, YouTube is leading the pack for podcast listening—it's got 28% of the market share, making it the top platform for podcasts, according to the New York Times.

And according to new data from Nielsen and Edison Research , younger listeners (ages 18-34) spent almost as much time with podcasts as they did with radio, at 37% versus 45%.

SHORT

READ

At one point search worked and then it didn't anmore. Here's the story of how Google Search died, and the people responsible for killing it.

On the other hand, apparently the Home Page is no longer dead.

The CEO of Axios thinks he can AI-proof the news. He's a smart guy so worth your time.

Smart and angry: A short rant—about the perils of anthropomorphizing AI by Zach Seward, head of editorial AI initiaves at the NYT .

While Zach argues we should treat generative A.I. models not as people but as tools, the People + AI research team at Google are making exactly the opposite point: "We need to help people move beyond thinking of AI as a “tool” that one learns to operate and control and toward thinking of it as a “collaborator,” and in the case of generative AI, as a “co-creator.”


That's it for this week. Another full plate, but if you consider it a buffet then you can come back for a second portion.

Wayfinder is made to travel. Feel free to share with friends and colleagues.

Ezra



?



David Fernández Quijada

Co-founder & Research Director @ South 180

6 个月

The study of Andrew McAfee is very disappointing, it's basically a desk research skewed towards what reinforces his views of the topic. There is no actual econometric analysis of the impact of GenAI nor even a fact-based and argued discussion of opposing views on key topics. Basically, after reading it I felt I had wasted my time.

回复

要查看或添加评论,请登录

社区洞察

其他会员也浏览了