Even if automation is creeping into all corners of our lives, at least we humans can still get together in real life to talk about it.
At the Algorithms, Automation, and News conference in Munich this week, some of journalism’s biggest brainiacs shared their research on everything from bot behavior to showing your work when it’s automated to reporting through the Internet of Things. Many of academics’ relevant papers will be published in a forthcoming issue of Digital Journalism. (Full list of presenters, panelists, and papers here.)
Here are some of the top tweets from the Munich mind-gathering:
The Associated Press’ director of information management Stuart Myles walked attendees through the AP’s process for making automation in news more transparent (hint: it includes automating transparency):
“How Can We Make Algorithmic News More Transparent?”
A proposed framework, presented at #AlgorithmicNews https://t.co/nECCr9JmOjhttps://t.co/nZTWGK26Z9
— Stuart Myles (@smyles) May 22, 2018
So @smyles from AP is suggesting four levels of algorithmic transparency:
• disclosure
• justification
• explanation
• reproduction (or reproducibility)#AutomatedNews— Axel Bruns (@snurb_dot_info) May 22, 2018
@HannaZoon, @jorgealveslino describing how they built “Charlotte”―an interview bot, that can conduct simple interviews with people, using journalist-defined questions and automatic follow-up questions.#AutomatedNewshttps://t.co/Qbot5NZsle@ifkw_lmu
— Neil Thurman (@neilthurman) May 22, 2018
Low-volume news users liked @abcnews‘s bots because they presented good news in a conversational, irreverent style, says @hfordsa – the bots humanised journalism, as it were. Which doesn’t reflect well on the language of standard mainstream journalism, really… #AutomatedNews
— Axel Bruns (@snurb_dot_info) May 22, 2018
The younger, more female & less educated users of @abcaustralia‘s newsbot liked its personal nature, how “news is presented in a fun, non-intimidating way.” So say@hfordsa & @dhutchman at https://t.co/Qbot5Oh3JO#AutomatedNews pic.twitter.com/CZHWQRyKoM
— Neil Thurman (@neilthurman) May 22, 2018
#Ethnographymatters! @hfordsa & @dhutchman presenting their research on #publicservicemedia and #chatbots: bots are seen as non-pretentious, positive, and helpful. It’s not the automation, it’s the journalistic tone which matters. #AutomatedNews pic.twitter.com/Pm5Ew7Gu9I
— Stefanie Sirén-Heikel (@MsSiren) May 22, 2018
Korea Press Foundation’s News Trust project coded 1000 news articles on various quality criteria & found, by their criteria, quality increased with length but only up to 1900 words.#AutomatedNewshttps://t.co/Qbot5NZsle@nettong@ifkw_lmu pic.twitter.com/mnlxhyXkcW
— Neil Thurman (@neilthurman) May 22, 2018
.@BerWeiss & @vicariSensor’s comparison between traditional journalistic storytelling and “sensortelling”, i.e. reporting that uses sensors embedded in the world and connected to the internet.#AutomatedNewshttps://t.co/Qbot5NZsle@ifkw_lmu pic.twitter.com/NgrWFcajEb
— Neil Thurman (@neilthurman) May 23, 2018
We aren’t there yet, but progress has been made on news automation or content planning @raufercon Study in 5 Swedish newsrooms showed that data processing and personalization are the most used digital features. Conclusions👇🏼Digital ecosystem👇🏼#AutomatedNews pic.twitter.com/83mMl0ngGP
— Juan-Carlos Molleda (@GlobalPRMolleda) May 22, 2018
Good questions for future research on automated journalism raised by @AljoshaKarim at #AutomatedNews conference here in Munich. pic.twitter.com/yiaQMaIcwp
— Seth C. Lewis (@SethCLewis) May 22, 2018
