domingo, 22 de abril de 2018

La tiranía del algoritmo


Los dejo con Eli Parisier explicando en una charla TED (California) la burbuja de filtros. La tiranía del algoritmo –que nos da lo que supone que nos gusta y borra lo que supone que no necesitamos– es hoy el peligro más tenebroso y el mayor desafío para las democracias occidentales. 

viernes, 13 de abril de 2018

World Press Photo 2018


La foto de la crisis en Venezuela, de Ronaldo Schemidt (AFP), que ganó el premio a la Foto del Año de la World Press Photo. Aquí todas las fotos ganadoras. Abajo los nominados.

Toby Melville (UK) Atentado en el puente del Big Ben.

Adam Ferguson (Australia) Chica bomba de Boko Haram

Ivor Prickett (Irlanda) La batalla por Mosul

Ivor Prickett (Irlanda) La batalla por Mosul

Patrick Brown (Australia) La crisis Rohingya

jueves, 5 de abril de 2018

Las noticias no son películas

Les paso la conversación por correo electrónico entre el GEN y Mark Little, de NevaLabs, tal como aparece en el blog del GEN (Global Editors Network).
Mark Little: No, we are not building a Netflix for News

The ‘Netflix for news’ analogy is often thrown around whenever there’s a new product that aims to shake up the way news is distributed. This has also been the case for Mark Little and Áine Kerr’s next venture, NevaLabs, so we wanted to find out what’s behind the gimmick. 
NevaLabs is a team of journalists, developers, and designers currently developing a personal assistant tool that aims to change the news reader experience from mindless scrolling into a daily routine. The final product will be controlled through an app, but delivered through multiple devices and platforms. User testing of the personal assistant prototype is currently underway. ‘We are ready for our assumptions to be challenged and our product to evolve’, said Little. 
In a conversation with the Global Editors Network, Mark Little, co-founder of NevaLabs, shrugs off the Netflix comparison and says he looks towards mindfulness apps for inspiration instead. He also gave us some insight about how his new media venture will handle user data, on what basis success will be measured, and why he won’t be trying to get people hooked on the service.
Mark Little and Áine Kerr
Food for thought from Little(compiled from an email exchange between Little and the Global Editors Network. Edited for clarity and brevity) 
Keeping it simple
We’re determined to build a radically improved user experience of news and information. A guiding operational principle for our developers and designers is therefore to make the right things easy. 
We’ve learned the value of bundling optimisation features together in user-friendly filters designed around time of day or state of mind. Artificial intelligence does a lot of the heavy lifting in simplifying the hierarchy of choices to be made by the user. 
Mindfulness apps and Instagram 
There are two analogies to keep in mind. First, the photography analogy, where mobile users now exercise creative control through amazing filters without ever getting bogged down in the technical complexity of the multiple enhancements that make them work.
Second, the rising popularity of health, fitness, and mindfulness apps that deliver a positive emotional charge around purpose-driven behaviour. That sense of agency over better outcomes is in the DNA of the products we’re developing.
 
Humans and machines 
Unlike the last generation of tech platform, we are putting the machine firmly in the service of the user. It’s not enough that the machine learns from the content the user clicks on, we want the machine to also give the ability to drill down into topics that match the user’s identity. 
As well as understanding the topics users want to proactively pursue, the machine should align to the daily habits of the user. It should offer audio and not video if the user is driving to work. How much time do they have on that commute? How does their mood differ on the journey home? 
We think of this as ‘purpose-driven personalisation’, a uniquely virtuous feedback loop between the human and the machine that helps the users identify and correct any imbalances or unintended biases. 
We are also experimenting with measures of geographic and gender diversity in a user’s feed, and would eventually like to build a recommender system to correct imbalances. 
News is not music. Or movies. 
We’ve grown increasingly wary of the Netflix or Spotify for News analogy. The data does suggest that news seekers — particularly younger users — are now more likely to pay for news because they have been conditioned to pay for music and movies. I also see evidence of a growing demand for access to multiple sources in personalised ‘playlists’ and I am inspired by the collaborative filtering that powers Spotify. 
But news is not music or movies. It doesn’t have the same shelf life. It plays a very different role in the lives of individuals and their society. News should not be unbundled from the authority of its original source. People want limits on the time they spend on news, not endless bingeing. 
So, no, NevaLabs is not building a Netflix for News. 
Re-imagining the public square 
Our system is designed to spot outlier topics and entities, helping users drill down into aspects of a story that are not immediately evident. Also, one of the most exciting discoveries we’ve made is the potential network effect of our system. We have come to see that we are creating a community of personal curators who are building highly curated reading lists and purposeful news habits that will be shared with other users in the network. This re-imagining of the ‘public square’ offers tremendous scope for serendipity, both in accidental discovery of new ideas, but also in collaborative filtering, which exposes users to different perspectives from similar people. 
How NevaLabs will change the news reader experience
Detached from advertising and social distribution 
One of the most liberating things about detaching news from advertising and social distribution is that we only care about data that is absolutely aligned with the user’s intentions. We’re completely in their service. We have no incentive to harvest, identify or store data that is not absolutely useful to the user. To back up best intentions, we’re actively seeking technology partners who can help store user data in personal silos the user can control rather than vast corporate databases. 
Emotion and misinformation 
We want to be the first media startup to build the concept of ‘trust metrics’ into our DNA. To begin, we want to provide clarity around the provenance of sources and content in the user’s feed. We want to alert users to content which has been contested by fact-checking networks (and show the substance of the challenge), and to monitor social signals which indicate manipulation or misinformation, such as extreme emotion. 
The strength of this approach is that it allows user to make informed decisions about the ongoing reliability of sources and adjust their preferences to reflect that. It also echoes the reality that not all contested content is untrue; sometimes fact-checks confirm the content. Our goal is always to put the user in conscious control of any outcome. 
We will likely base this feature on an open API developed by a technology partner. We’ve been experimenting with the claim review process which Google developed (and until recently had been testing publicly) and we’re excited by the evolving work of the Credibility Coalition and other open-source groups like First Draft News and the Trust Project. 
Main metric of success: Time well spent 
Non-profit innovation is necessary but not sufficient if we’re to rebuild the foundation of journalism in a digital age. We need a constellation of startups delivering constant, sustainable innovation in excellent user experience of news and information. The goal is something I like to call an ‘economy of trust’ for journalism. 
We are inspired by research that shows the apps people are happiest with are those they spend small amounts of quality time with, like health, fitness, and mindfulness apps.
At NevaLabs, the metric of the system we are building is time well spent: how much value can we deliver in the least amount of time. Our business model means we have no interest in keeping the user addicted to our platform.
 
Thanks Mark!