View this in a web browser Add to your address book
EBU Banner

March 2019


A lot happened these past few weeks. We held our Data in the Newsroom workshop in Munich. The Council of Europe also organized its own major event on governing AI. Then, the World Wide Web turned 30 in March and the Internet turns 50 soon



Data in the Newsroom workshop 26-27 Feb, Munich
At the 'Data in the Newsroom' workshop in February, participants explored how data, AI and ML can be used to improve workflows for content production, the fight against misinformation, personalization of content, data journalism, and more.

The workshop also explored the strategic and ethical consequences for PSM: is journalism needed in this era of social media? And how do we ensure algorithms respect journalistic values and ethics? 

We'd like to thank all participants for coming and the IRT for kindly hosting us. See you in Manchester, 26-27 November, for the next Workshop!.


Council of Europe holds AI conference - Following its 2-day 'Governing the Gamechanger' conference, and recognizing the potential of AI for both good and bad, the Council of Europe has produced a series of recommendations. It notably emphasizes the importance of ensuring algorithmic transparency, human-centric AI designs, and the necessity of ensuring that AIs do not violate human rights principles and legislation.

The World Wide Web hits 30 - 30 years ago, Sir Tim Berners-Lee invented the basics of the World Wide Web, changing our lives forever. Berners-Lee and his invention were celebrated on March 12 at CERN in Geneva, and (thanks to the Web) all over the world.  

An AI that auto-generates believable text - OpenAI has created 'GPT-2', a Machine Learning model trained to suggest the next word in a text, based the what came before. The model is now capable of generating believable text sequences. Wary of the potential for misuse, the creators haven't released the model in full. 

An AI that detects auto-generated text - Researchers at Harvard and MIT have create the 'Giant Language model Test Room' (GLTR), an algorithm that reverse engineers text sequences using common language models. You can use it to test whether a given text sequence was likely written by a bot.


Fun fact: hosts a large collection of pictures of people that do not exist – because all of the (very realistic looking) pictures listed there were randomly generated by an AI! You can also find out more about the technology used to generate these pictures.


Visit to see all EBU upcoming events, plus presentations and videos from past events.


Do you have any suggestions for future content? What are the topics that you would like to see most? Send your feedback to