For six months, Media City Bergen and four affiliate organisations collaborated with newsrooms accepted to ascertain how AI can ability storytelling. Here, we allotment our adventure and our findings.
Media City Bergen is a Media Addition Cluster based in Norway. This spring, we were accepted into the 2021 JournalismAI Collab Challenges. We spent the abutting six months experimenting our way to next-generation AI-assisted storytelling with acclaimed newsrooms such as The Guardian, Deutsche Welle, Bayerischer Rundfunk, Science Media Center Germany, and Sveriges Radio.
As a Media Addition Cluster, Media City Bergen is a bit altered from our collab peers. We accept been a collab aural the collab, with four abstracted affiliate organisations accommodating beneath the Media City Bergen umbrella. The organisations are the banking newspaper Dagens Næringsliv, the bounded newspapers Fædrelandsvennen and Stavanger Aftenblad, and the AI-based-fact-checking-company Factiverse. It was the aboriginal time a bunch of organisations alternate in the collab Challenges, which fabricated us alike added aflame about the adventure ahead.
How ability we use modular journalism and AI to accumulate new storytelling formats?
Our claiming was this: The aggregation will analyze opportunities to acquaint journalistic acceptance in altered means to altered bodies in altered contexts. A specific focus will be on developing new storytelling formats to ability currently underserved audiences. How ability we use modular journalism and AI to accumulate new storytelling formats?
Phrased differently: How can we anatomy our journalism in abate pieces that can be stitched calm in altered means at alone lengths to serve several ambition audiences in means that bigger bell with them than accepted journalism does?
It’s a appealing ample scope, so our aboriginal footfall was all about absorption the focus. The accomplished organisers at Polis – the journalism think-tank at the London School of Economics and Political Science and advisers from Clwstwr and BBC Annual Labs led us through a architecture cerebration process. We looked for so-called underserved audiences – readers, watchers, and admirers that are not currently served abundantly by annual media. There are abundant underserved admirers groups out there. We absitively to focus on young readers – accurately those amid 18 to 24 (Generation Z) and 25 to 35 years old (Generation Y aka Millennials). These segments are about beneath affianced by acceptable annual acceptance than earlier audiences, so best annual organisations assignment adamantine on award bigger means to serve them.
From our aboriginal concrete aggregation affair in Arendal in August 2021. From larboard to right: Ketil Moland Olsen (Media City Bergen), Frode Nordbø (Fædrelandsvennen), Gaute Kokkvold (Factiverse), and Mailinn Mersland (Fædrelandsvennen).
So what’s the problem? Why are these consumers not engaged? Let’s alpha with the archetypal way adolescent readers currently absorb news. As declared in a recent Reuters Institute report:
While those over 35 are acceptable to aboriginal go anon to a annual armpit via an app or the adaptable browser (39%), Gen Z are added acceptable to about-face to amusing media and messaging apps (57%). In added words, annual brands are beneath important for this accumulation than for over 35s. Gen Y are about in the middle, with 43% accepting their annual via amusing media and messaging apps and 33% directly.
Younger readers await added heavily on amusing media to amend themselves, which poses a claiming – not alone for the annual organisations but potentially additionally for the sustainability of our avant-garde democracy. Amusing media is about tailored to the reader’s tastes and does not necessarily accommodate a counterbalanced appearance of reality.
What does this beggarly for the annual brands? The Reuters Institute address continues:
Insights from our agenda tracking of annual users’ adaptable burning reiterate that annual brands comedy a actual baby role in adolescent people’s lives. Best smartphone time was taken up by amusing arrangement apps, internet browsers, podcasts, mail, and movie/music alive accessories – followed by dating apps, maps, and carriage applications. Adolescent bodies accept a actual low beginning for apps that don’t accommodate a abundant experience, while they bulk casework that are accordant and advantageous at all times. No annual app was aural the top 25 apps acclimated by all the respondents in the study, admitting Instagram was the appliance begin on about all phones with the accomplished use in agreement of circadian annual used.
There is some consolation to be begin in the report:
This does not beggarly that acceptable brands are not admired by adolescent consumers. Best do accept an ‘anchor annual brand’ that they will about-face to aback a above adventure break and needs acceptance – in our qualitative assay abstraction this was about the BBC or Guardian in the UK, and CNN or the New York Times in the US. The best of this cast is generally heavily afflicted by aboriginal affectionate access but the architecture is about consistently digital.
To achieve on a absolute note, there seems to be a abeyant for extensive the adolescent ancestors directly, but we charge to do some abundant appropriation to acquisition out what works for them.
Using adolescent capacity and sources in the storytelling ability calmly access agreeable burning in the adolescent ambition group.
Now we had our ambition group. The abutting footfall was analysing bazaar abstracts calm by newsrooms in our team. Aggregated qualitative and quantitative abstracts appropriate that adolescent bodies are absorbed in abundant of the aforementioned as earlier readers. Allotment of the challenge, it turns out, is accompanying to the persons in the annual stories. At atomic for some of our newsrooms, the subjects and sources in their advertisement are about essentially earlier than the ambition accumulation they are disturbing to reach. According to the analysis, application adolescent capacity and sources in the storytelling ability calmly access agreeable burning in the adolescent ambition group.
Furthermore, surveys and interviews from our newsrooms’ centralized assay announce that adolescent bodies appetite to learn more from journalism. As an experiment, one of our accommodating annual organisations attempted to emphasise acquirements credibility and distilled analytical takeaways in the text. The ambition was to advice the readers accept what ability they should get by annual the article. And anticipation from the acknowledgment afterwards the experiment, highlighting the acquirements credibility and takeaways seems to be a way to access both annual and apperception amid the young.
An absorbing absurdity alike during the discussions in our assay phase. We apperceive that editors continuously optimise the angling of acceptance and the agreeable mix of advanced pages for the boilerplate customer persona. It increases the readership, but it additionally optimises agreeable for the boilerplate clairvoyant – who about belongs to the earlier clairvoyant group.
What happens, then, aback we broadcast an commodity that targets adolescent audiences accurately and doesn’t appoint earlier readers at all? We get low abstracts and low afterimage on the advanced folio – alike admitting this adventure ability finer appoint the adolescent readers we are aggravating so adamantine to reach. We can break some of this botheration with personalisation, but that comes with its own set of challenges that we charge to mitigate.
Moreover, the group’s centralized testing, commutual with alien research, indicates that the format of the agreeable plays a cogent role aback adolescent bodies appetite to amend themselves on the news. And here, the accent of accounting argument cannot be overstated. In addition, a aggregate of sound, image, and argument customised for adaptable platforms seems to assignment well. As this happens to be the architecture of Instagram and Snapchat, this is hardly surprising.
Let’s booty a moment to sum up our adventure so far. We accept articular an underserved group – the adolescent consumers – and we accept abstruse absolutely a bit about why they are currently underserved. Our abutting step: To acquisition means of better confined them. And as the actual name of the 2021 JournalismAI Collab Challenges suggest: AI should be allotment of the mix.
Anyone complex in artefact development knows that it’s all about creating solutions that solve problems rather than creating solutions that look for problems. And this is decidedly hard. For starters, it’s almost aboveboard to dream up confusing and potentially world-changing solutions for approaching ancestors – which would additionally crave massive changes in the accustomed journalistic workflows. On the added hand, it’s accessible to abatement into the allurement of over-engineering the acquaintance for the consumer. Do they appetite to angular back, angular forward? That depends on the context. And then, it’s the endless added variables we charge to booty into account. Striking a antithesis is challenging.
One of the team’s aboriginal account proposed alteration the way we address acceptance from the basal up. Journalists would access tiny atoms of advice accomplished with metadata into a specialised CMS and afresh augment the consistent abstracts structures into a apparatus acquirements algorithm that would stitch the adventure calm in a custom-tailored way for a specific customer in the accepting end. As our group’s assay indicates, adolescent bodies appetite beneath argument and added visuals. The appropriate access could absolutely be a footfall in this direction, customising agreeable to the accepting end afterwards manually abstraction out several versions of the aforementioned story. Accustomed that we alone had a brace of months to accomplish our eyes a absoluteness (the collab had a borderline in November), we bound realised that this abstraction was out of ambit for now and connected baking bottomward the bulk idea. Still, account from this appearance were beneficial, and bits accept begin their way to our accepted prototype. (Our collab peers Deutsche Welle, Il Sole 24 Ore, Maharat Foundation, and Clwstwr created a accompanying concept. You can analysis out their abstraction at their project website.)
We connected tinkering with both the botheration and our brainstormed band-aid ideas, simplifying in both ends as we progressed. And as it angry out, sometimes agitative addition comes in increments rather than in the behemothic leaps of cogent disruptions. Aboriginal this fall, we had above the abstraction bottomward to an aspect we could assignment with and a proposed band-aid that we could realistically body and analysis during the accustomed timeframe.
Our botheration assumption: News journalism generally takes bulk concepts for granted. Adolescent annual consumers generally abridgement ability about these concepts. Hence, the defective ability becomes a barrier to access for the accustomed consumers.
Our band-aid idea: By highlighting words and concepts we accept to be difficult and appropriately embedding baby allegorical actuality boxes in the text, we ability arch the bulk concepts ability gap for consumers who charge it.
We hypothesised that we could adorn absolute journalism with added advice by application simple actuality boxes and called our idea Micro Actuality Box. Nice name, but would it survive the antecedent acquaintance with users?
We hypothesised that we could adorn absolute journalism with added advice by application simple actuality boxes and called our idea Micro Actuality Box. Nice name, but would it survive the antecedent acquaintance with users?
To acquisition out, we set out to body a prototype. Luckily, we had several accomplished developers on the team. One of them, Frode Nordbø from Fædrelandsvennen, conjured up a alive audience aloof hours afterwards we nailed the antecedent idea. He alike wrote a simple algorithm that ranked all the words in the Norwegian accent from best to atomic acclimated – as an aboriginal way of award words to highlight. We could now run argument accessories through Frode’s algorithm, and it would automatically tag the best difficult words – in this case, atomic acclimated words in the Norwegian accent – for us. Some of the words were somewhat hasty and out of context, but we absitively it would be acceptable to prove our concept.
Next, users should be able to bang on the accent words to see their definitions. Instead of implementing the classical pop-up box, accustomed from Wikipedia, Frode fabricated a able twist. He advised an in-line actuality box that, aback clicked, appeared amid two paragraphs in the article. We now had the words, and we had the micro actuality boxes. But area to acquisition the agreeable for the actuality box? For the prototype, we absitively to cull the definitions beeline from Wikipedia.
Our antecedent Micro Actuality Box prototype. Actuality boxes amid the paragraphs aggrandize into afterimage aback the user clicks on a accent keyword or byword in the text.
While we were active developing our prototype, we realised that our proposed band-aid adequate on a hardly cool foundation of assumptions. We bare to analysis added with absolute consumers from the appropriate ambition groups. So while Frode connected alive on the prototype, added aggregation associates formed on user research. We attempted to accouterment the best analytical acceptance arch first: Do adolescent readers abridgement ability about bulk concepts in annual articles?
To acquisition out, Dagens Næringsliv conducted a low-tech agreement for us. They sampled three accessories from their publication, printed them, and handed the copies out to a console of adolescent subscribers. Then, they asked the participants to apprehend the accessories with pen in hand, highlighting any difficult words or concepts they came across. We were afraid by the results. About none of the participants had accent any words or concepts at all. One of the accessories had a few accent words, but the bodies who had fabricated highlights commented that a acceptable account came after in the text.
If we were to do this analysis again, we would probably quiz the readers about apperception rather than ask them.
Had we headed beeline bottomward the amiss track? Strictly speaking, we could accept taken these after-effects as absolute affidavit that we were analytic the amiss botheration with our Micro Actuality Box concept. However, we realised that we ability accept been allurement inappropriate questions – or rather allurement questions to the amiss persons. The analysis console participants were all subscribers of Dagens Næringsliv, and – as advantageous barter – we would apprehend them to acquisition the publication’s agreeable aural reach. On the added hand, it’s acceptable that they were not adumbrative of the boilerplate being in Generation Y or Z. This angry out to be one of several capital acquirements credibility for us. If we were to do this analysis again, we would probably quiz the readers about apperception rather than ask them.
Meanwhile, artist and frontend developer Alexander Prestmo from Dagens Næringsliv abutting armament with Frode on the project. He implemented the aboriginal abstraction with DN’s attending and feel and added some accomplished new features, such as a user-selectable bulk of keyword extraction.
Now, we were accessible to do the aboriginal real-world user analysis of the prototype. We recruited a scattering of acceptance amid 20 and 29 years old, arrive them to Dagens Næringsliv’s offices, and let them do an unsupervised analysis drive of our solution. It was a bit of a nerve-wracking experience, but that activity bound afflicted to abatement aback we realised that the analysis pilots allegedly accepted our Micro Actuality Box concept. Indeed, some wrinkles bare ironing, but the tests added our acceptance that we were assimilate something. Actuality are some of the comments we aggregate from the analysis participants:
I anticipate this affection is advantageous aback annual accessories with complicated words.
It’s benign if you are a non-native apostle of the language.
It ability force me to apprehend added of the text.
It should focus on the “hard words”.
You should accompaniment why you highlight assertive words.
It was affective to accept this feedback, but there are still a lot of questions to answer. The accustomed abutting footfall would be to analysis the abstraction quantitatively with a beyond sample of annual consumers. We couldn’t do this in time for the collab deadline, but don’t be afraid if you see Micro Actuality Boxes actualization in your favourite online advertisement soon.
Eager to test the prototype? You’ll acquisition it here.
A adolescent user is testing our ancestor in Dagens Næringsliv’s offices. Gaute Kokkvold from Factiverse is observing.
No JournalismAI Collab Challenges without AI as allotment of the solution. And this is area Media City Bergen-member Factiverse and their assembly Maria Amelie and Gaute Kokkvold comes into the picture. Factiverse is an organisation that develops accoutrement that automate assay and apprehension of misinformation with cutting-edge AI and Accustomed Accent Processing. They accept developed the apparatus acquirements agent and APIs during the collab that will abstract the arduous words and concepts in annual articles. The agent will additionally attending up keywords and abstraction definitions from third-party sources such as Wikipedia and added online encyclopedias. It’s still in the aboriginal stages of development, but the agent can already analyze amid people, places and organisations. Furthermore, Factiverse is developing a journalist-facing frontend, area the accent words can be absolute and modified. The plan is to accommodate this frontend with altered newsroom CMSes.
Highlighting the right words for the right readers at the right time is one of the cogent challenges advanced of us.
Highlighting the right words for the right readers at the right time is one of the cogent challenges advanced of us. Personalisation seems like an assured basic here, and acknowledgment from readers (through tracking what micro actuality boxes are opened and not opened) can be an able way of training the apparatus acquirements algorithm that decides what keywords to highlight. We are attractive advanced to accepting to this date of the project.
As we address this, the project’s developers are busily abutting the solution’s front-end and back-end components. Aback the pieces are puzzled together, we will be able to analysis the complete, automatic workflow. We are optimistic that the abstracts conducted during the 2021 JournalismAI Collab Challenges will accomplish it into the newsrooms’ artefact in one way or another.
We achievement that our Collab agreement is the aboriginal of several accomplish to explain the annual in a way that works for adolescent audiences.
Crunch time multitasking in Oslo. While accommodating in the annual collab cachet meeting, Gaute Kokkvold (Factiverse) and Alexander Prestmo (Dagens Næringsliv) accomplish some adjustments to the prototype.
When innovating in the circle amid journalism and acid bend technology, it’s accessible to anticipate too big too fast. While alive on the Micro Actuality Box project, we accept apparent a lot of possibilities for accretion the concept. “Why don’t we actualize a CMS for actuality boxes area we can administer the argument and stitch the boxes calm as absolute timelines? And while we are at it, why don’t we advance a apparatus acquirements algorithm that writes the argument in the actuality boxes for us and creates customised versions for the assorted ambition groups?”
These ability all be abundant ideas, but we charge to apprentice to airing afore we can run. By acceptance the addle pieces individually, it’s added aboveboard and beneath chancy to aggrandize the bulk concept. AI will acceptable be an abundantly added cogent allotment of the mix as we accumulate developing the idea. Apparatus acquirements is already anecdotic the key passages of the accessories and will do so added added absolute as we alternation the algorithm. After on, apparatus acquirements ability additionally be autograph or adapting the argument we affectation in the actuality boxes.
When acclimated optimally, apparatus acquirements ability enhance journalism significantly. But it’s not a abracadabra bullet. And no, bots will not booty over the journalistic profession anytime anon and accomplish us bodies redundant. We can, however, apprehend them to be added well-functioning “colleagues” shortly—colleagues afterwards a faculty of humour and cultural knowledge. The best cogent abeyant of AI in journalism is acceptable to be begin in the circle amid bodies and machines. Together, we ability aloof change the apple – a tiny bit at a time.
But it consistently starts with the users (in this case, both journalists and consumers) and their needs. Let’s not balloon that.
Our best accepted adaptation of the abstraction at the time of publishing. Agenda the new affection that lets users baddest their desired amount of actuality box enrichment.
The 2021 JournalismAI Collab Challenges has been a altered experience. For the Media City Bergen team, it’s been about acquirements to coact as a accumulation (the Media City Bergen participants) aural the accumulation (the EMEA collab team) aural the accumulation (the all-around collab team) – a Matryoshka baby of altered layers, in added words. We accept tried, bootless and abstruse together, but the best agitative allotment has been acquirements calm as an all-embracing ecosystem. It’s been alarming to see what the added teams accept created. At the aforementioned time, it’s absorbing to realise that all of us face abounding of the aforementioned challenges apart of differing cultures and our concrete locations in the world.
We would like to booty this befalling to acknowledge all our 2021 JournalismAI Collab Challenges peers. It’s been alarming to apprentice from the global community. Midway through the project, we begin accepted arena with Bayerischer Rundfunk and Science Media Center Germany, who formed on a activity with similarities to ours. We appetite to acknowledge you attentive for a affable and abounding barter of ideas, problems, and solutions. A appropriate acknowledgment go to Marco Lehner, Cécile Schneider, Rebecca Ciesielski from BR, and Meik Bittkowski from Science Media Center Germany. Accomplish abiding to check out their activity website.
From our own teams, we appetite to acknowledge Anders Hamre Kulien and Siri Andresen from Fædrelandsvennen, Frode Buanes, Julie Lundgren and Petter Winther from Dagens Næringsliv, Lars Helle and Elin Stueland from Stavanger Aftenblad, and Marianne Fjellhaug and Charlotte Vindenæs from Media City Bergen for abundant abaft the scenes-work.
Finally, a big acknowledge you goes to the collab organisers: Mattia Peretti and his aggregation at Polis – the journalism think-tank at the London School of Economics and Political Science, and our advisers Shirish Kulkarni from Cardiff Clwstwr and David Caswell and Jane Birch from the BBC. We are beholden for the befalling and achievement to be accommodating with you all afresh in the future.
The Micro Actuality Box Team: Frode Nordbø (Fædrelandsvennen), Mailinn Mersland (Fædrelandsvennen), Rolf Frøyland (Stavanger Aftenblad), Maria Amelie (Factiverse), Gaute Kokkvold (Factiverse), Ingeborg Volan (Dagens Næringsliv), Alexander Prestmo (Dagens Næringsliv ), Anne Jacobsen (Media City Bergen ), and Ketil Moland Olsen (Media City Bergen) .
We presented our assignment at the JournalismAI Festival. You can watch it below.
That’s all, folks! But afore you go, accomplish abiding to analysis out the publications and websites of our collab peers. It’s a massive bulk of hard-earned acquaint to be abstruse in there, and you can accept it all for free:
The commodity was aboriginal appear on Media City Bergen’s website.
This activity is allotment of the 2021 JournalismAI Collab Challenges, a all-around action that brings calm media organizations to analyze avant-garde solutions to advance journalism via the use of AI technologies.
It was developed as allotment of the EMEA accomplice of the Collab Challenges that focused on “How ability we use modular journalism and AI to accumulate new storytelling formats and ability currently underserved audiences?” with the abutment of BBC Annual Labs and Clwstwr.
JournalismAI is a activity of Polis – the journalism think-tank at the London School of Economics and Political Science – and it’s sponsored by the Google Annual Initiative. If you appetite to apperceive added about the Collab Challenges and added JournalismAI activities, sign up for the newsletter or get in blow with the aggregation via hello@journalismai.info.
How To Write A Glossary For A Report – How To Write A Glossary For A Report
| Welcome for you to our weblog, with this moment We’ll explain to you with regards to How To Delete Instagram Account. And after this, this is actually the 1st picture:
What about graphic earlier mentioned? will be which incredible???. if you feel consequently, I’l t demonstrate a few picture once again down below:
So, if you like to secure all of these great pictures about (How To Write A Glossary For A Report), simply click save icon to store these pics for your computer. They’re ready for save, if you appreciate and want to have it, simply click save badge in the article, and it will be immediately saved to your laptop computer.} Finally if you would like receive unique and recent graphic related to (How To Write A Glossary For A Report), please follow us on google plus or book mark this page, we attempt our best to present you daily up grade with fresh and new shots. Hope you enjoy keeping here. For most up-dates and recent news about (How To Write A Glossary For A Report) images, please kindly follow us on twitter, path, Instagram and google plus, or you mark this page on bookmark area, We attempt to present you up grade periodically with fresh and new shots, enjoy your searching, and find the best for you.
Thanks for visiting our site, articleabove (How To Write A Glossary For A Report) published . At this time we’re delighted to declare that we have discovered an incrediblyinteresting contentto be pointed out, namely (How To Write A Glossary For A Report) Many people searching for information about(How To Write A Glossary For A Report) and definitely one of them is you, is not it?
/Getty_glossary-115709768-5701c8135f9b586195354c69.jpg "Definition and Examples of a Glossary in a Book")
