Notices by DW Innovation (dw_innovation@mastodon.social)

A big shout-out to the #ARD & #ZDF family on this platform:

@3sat
@BR24
@DasWissen
@ki_und_koffein
@NDR
@swr3
@tagesschau
@ZDF
@zdfmagazin

Dear Fediverse people,

It's time to say thank you again. 🙏

We started this account 2 years ago & we've recently hit 7k followers. 🎉

What's more important than this standard metric: We've received a lot of insightful comments & questions (& a few less insightful ones 😄).

The point is: This feels like a great platform for an R&D team working at the intersection of #media #technology & #journalism.

We''re here to stay & we're trying to convince others to join as well.

Thx again & cheers!

Potentially powerful political storytelling, enabled by media technology:

The #UN recently commissioned a #VR experience that gives viewers "a real, tangible sense of what peace keepers do" in South #Sudan – "which often gets lost in discussions and presentations."

More in this DW video report:

https://www.dw.com/en/virtual-peacekeeping-empathy-in-action/video-74228103

"The fact that Tilly Norwood and the company behind ChatGPT are simultaneously engendering such controversy is not a coincidence: This is an existential moment for human-created entertainment as we know it. If actors, talent bookers, and studio executives cannot hold the line now, at this very moment, the battle to preserve the humanity inherent to art will be irredeemably set back."

(via @slate)

https://slate.com/technology/2025/10/openai-sam-altman-sora-ai-actor-tilly-norwood-hollywood.html

#AI #movies #art

"Imagine a world where scientific experiments yielded different results every time you ran them, even with identical setups. (...) In most fields, such inconsistencies would be unacceptable. Yet, in the (.) world of large language models (LLMs), we’ve often accepted a curious paradox: the same prompt, run multiple times, can produce entirely different outputs."

Anthony Laneau for @hackernoon:

https://hackernoon.com/the-unseen-variable-why-your-llm-gives-different-answers-and-how-we-can-fix-it

#AI #LLM # nondeterminism

Dear followers,

It's time to take another social media break.

We'll be back on August 14th.

Until then, check out our blog archive, our collection of tools and prototypes – and DW's other services.

Stay safe, have a good summer, see you later! 👋

📸 (by @lxplm: The Rhine, the Kennedy Bridge and a part of Bonn's skyline, photographed from a rocky beach in Castell.

NewsGuard, a company that aims to help identify reliable news sources, has stopped using the labels 'misinformation' and 'disinformation' as they were politicized and partisan. They now explain whether a claim is '[...] explicitly false, AI-generated, unsubstantiated or manipulated'.

What do you think?

By Sophie Culpepper for @niemanlab.

👇
https://www.niemanlab.org/2025/07/language-that-once-clarified-is-now-obscuring-newsguard-retires-the-labels-misinformation-and-disinformation/

#Misinformation #Disinformation

Data scientists at York St. John University have published a new tool for identifying fake images: Pixelator v2. The software can detect alternations as small as one pixel.

Article by Alex Scroxton for Computer Weekly.

👇
https://www.computerweekly.com/news/366616231/Scientists-demonstrate-Pixelator-deepfake-image-verification-tool

#Verification #Images

Vertical #video is still a massive trend, like it or not. And at some point, even hardcore word people and those dedicated to old school web #journalism will have to produce one of those clips. A new #app called Sophiana might be a useful #tool in this case.

The @niemanlab has taken a closer look at the thing – and also talked to its creator, Sophias Smith Galer:

https://www.niemanlab.org/2025/06/this-new-app-wants-to-help-prose-loving-journalists-make-the-leap-into-vertical-video/

Check out DWA's video series on "#AI for media", discussing #disinformation and #MediaLiteracy.

Topics range from "How Conversational AI is changing #journalism" to "AI assisted #FactChecking in Arabic” – with a range of experts and researchers.

(Yes, it's on YouTube, because).

https://www.youtube.com/playlist?list=PLWRGkKg3JC1iQbZab_hXIcKG1QL6tfefv

(1/2)

#AI #intimicacy #information

"We're seeing the emergence of what I call the 'AI intimacy dividend'. Value is created by this new and fascinating willingness people seem to have to open up to conversational AI interfaces in ways not possible before, and it’s about to transform how we engage with news and information."

Yes, we know: #RSS isn't exactly cutting edge. 🙂

However, in a world of social media noise and click bait, promoting tried-and-true web services can be smart and innovative, too. 😎

So here's a collection of #DW feeds you can directly tap into via your favorite traditional #news aggregator:

https://corporate.dw.com/en/rss-feeds/a-68693346

Russia's war on #Ukraine, visualized in a collection of annotated charts:

- Ukrainian territory controlled by Russia
- missiles and artillery fire
- civilian deaths and injuries
- damaged civilian structures
- internally displaced people and Ukrainian refugees abroad

https://www.dw.com/en/human-toll-of-russias-war-in-ukraine-explained-in-graphics/a-71495389

(incl. code and data repository, via DW Data)

#DDJ #DataViz

Everybody is talking about #AI – but the focus is usually on the western world and China. Here's a video report on how the technology is changing societies and economies in #Africa:

https://www.dw.com/en/ai-in-africa-how-is-the-continent-tapping-into-new-tech/video-71594268

Lost in the #GenAI and #LLM jungle?

Here's a concise overview of the most important models and what you can do with them:

https://techcrunch.com/2025/02/17/the-hottest-ai-models-what-they-do-and-how-to-use-them/

(via @TechCrunch)

Stewart Lee, commentator at the @guardian, declares that the dark age of social media (and information in general) is upon us – and opts to log off.

He has a couple of very strong arguments:

https://www.theguardian.com/media/2025/jan/19/information-dark-age-elon-musk-donald-trump-jeff-bezos

However, let's not forget there are always democratic structures and alternative platforms we can use and foster – like Mastodon and the Fediverse.

#disinfo #propaganda #socialmedia

New year, new project: MOSAIC is all about access to vast European #video #archives, multi-language #content #streaming, tools for content creation/sharing, and #AI.

Mirko Lorenz has more detailed info on our blog.

https://innovation.dw.com/articles/mosaic-europe-video-streaming-platform

Castopod is an open-source podcast hosting platform that integrates with the Fediverse and has a strong focus on data sovereignty (your content, your audience, your analytics):

https://docs.castopod.org/main/en

#fediverse #podcasting

Hot off the #OSINT blog press:

A comparison of digital, partly AI-driven #tools for #geolocation.

Our Ruben Bouwmeester and @twone2 have taken a closer look at Overpass Turbo, GeoGuessr GPT, GeoSpy, Earthkit – and DW's own SPOT application.

A key insight: geolocation tools aren't perfect yet, but they're becoming more efficient and powerful every day.

https://innovation.dw.com/articles/geolocation-verification-tools-comparison

Hi everyone!

Time for a quick meta update:

We've hit 6k followers – which is amazing. More importantly, we still see a substantial number of interactions, and we'd like to say thank you for that! Please understand that we can't always read everything or even address all comments.

As for technical news, you should also have access to DW Innovation via Bluesky now. That's courtesy of @bsky.brid.gy (and @newsmast who pointed out this option).

Ok. Back to R&D work. See you later.