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


May 2019
In this edition, we explore the need for public service recommendation algorithms; the European Commission's recently published ethical AI guidelines; an app that lets you see only positive tweets; an upcoming Metadata and AI workshop; some thoughts on why data science teams might need generalists not, specialists; and plenty more to read and do.


Ben Fields, speaking at our AI in Media workshop
Recommendation algorithms offer an efficient and automated way to get users closer to content that matches their preferences. But the technology can play into a person’s biases if applied carelessly. How can PSM balance efficiency with the public interest?

We spoke with Ben Fields, lead data scientist at BBC News, about public service recommendation algorithms, what sort of values could form the basis of better algorithms, the impact of AI in news, 'deepfakes', and the challenge to build 'trust in news'.

The interview is publically accessible. EBU Members can also stream Ben's presentation at an EBU AI workshop last year.


EC publishes ethical AI guidelines - On Digital Day 2019, the EC pushed into the spotlight their guidelines for trustworthy AI. You can provide feedback during the AI guidelines pilot phase and exchange best practice on their dedicated forum.

Save the date! Metadata and AI workshop (11-13 June) - The annual EBU Metadata Developer Network (MDN) workshop will be held in Geneva. It'll cover everything related to metadata and AI in the broadcast world. An essential workshop for all developers!

AI beats E-sports world champions - OpenAIFive, an AI gaming machine, recently bested the world champions of DOTA 2, team OG. How did it do it? It trained, on average, a whopping 250 years' worth of simulated gameplay per day, for 10 months straight.

Ethics of AI at WSIS 2019 - The World Summit on the Information Society debated AI governance and AI ethics. It highlighted the need for a multi-stakeholder approach and discussed the possible need for a global framework.

DataSci awards deadline approaching -The application deadline for the 4th annual European DataSci & AI awards is at the end of May. If you have a great project you might want to check out their 10 different award categories.

Open-source data-visualization tool - Plotly has published Plotly Express, a data visualization tool using Python. It uses an open-source MIT licence and is freely available, even for commercial use.

Alan Turing Award goes to Machine Learning pioneers - the Association for Computing Machinery (ACM) has awarded the Alan M. Turing prize to Yoshua Bengio, Yann LeCun and Geoffrey E Hinton. Without their contributions to Machine Learning, Deep Learning and Artificial Neural Networks, the field of AI would far less advanced than it is today, the ACM says.


The MIT has developed Affective Network, a chrome extension that colour-codes and filters twitter posts by their emotional content. You can see how much positive, negative or neutral tweets fill up your feed and block the downers. How this kind of filter would affect society if widely used, is an open question.


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


Hi! You've reached the bottom. While you're here, how about sending your suggestions and feedback to