Talk:Strategy/Wikimedia movement/2018-20/Working Groups/Partnerships

We invite everyone to have a look at the existing documentation of the Working Groups, and then add comments, additional input or share concerns via this talk page. Your comments will be taken into consideration by the respective Working Groups.

Partnerships to beat back disinformation and take back the internet

The Wikimedia Foundation should move out ahead of the disinformation storm by partnering with internet giants who now understand their platforms are being manipulated by propaganda. Together, they all should share resources, ideas, and mostly talent—engineers, data scientists, trust and safety, community development, designers—to fortify the knowledge commons. The information wars are gaining ground, but there is still time. We'll need vision, moral clarity, people, bots, altruistic algorithms... and true partnership to take back the infosphere.

This is a partnership that should be forged at the highest levels across their organizations, not relegated to a department or a working committee, such as this one.

Winter is coming. We must fight our way back to springtime on the internet.

Examples of collaboration pathways:

1. Get the partners to lend Wikimedia engineers, like FB did for HHVM.
2. Give Aaron Halfaker engineers from FB, Google, and YouTube to build altruistic algorithms that intelligently fight bad faith edits without destroying the social experience.
3. Gather the elves to fight the growing army of trolls. It's a working group across organizations (I first learned of the Internet Research Agency from Philippe Beaudette in 2015 when he was working in Trust and Safety. He posted this in June of 2015. He also knows all the pockets of Reddit and 4chan).
4. Let us all take a lesson from Brandon Harris. We must learn to "design for evil" not just out of hope and naivete: Designing for evil.
5. Partner with Danny Hillis. He's building The Underlay. He wants it open and he wants it free. "The public knowledge that will reside in the Underlay is currently presented in mostly human readable overlays, such as publications, charts, and dynamic user interfaces. It is also represented in diverse databases in various incompatible formats. The Underlay will expose and collect the knowledge in those presentations into a standardized machine-readable form, with a record of its provenance. Separating the underlying knowledge from the overlay of presentation will help machines analyze its usefulness and veracity. It will allow the knowledge to be recombined and re-presented for other purposes. This will save lifetimes of duplicated effort and accelerate human progress. Through our machines, we will take advantage of the Underlay’s record of assertions and their provenance to help judge what is true." Since he built the knowledge graph and sold it to Google, he's got the vision and the technical chops. He also has the altruism and the ethics.
6. FB has already admitted that they need Wikipedia.
7. So has YouTube.
8. Bring in Tim Berners Lee. He's heartbroken and could use some ideological companions.

A broader coalition for the future.

While we get to work on beating back disinformation (at least a five-year collaboration before stabilization), there is a broader coalition with a revenue model to be built. If Wikimedians fear the influence of this money... fine, place it in the endowment with an agreement of how it could be dispensed to the Foundation YoY.

We document knowledge and give it away.

For the last 16 years, The Wikimedia projects have hosted communities of critical thinkers around the world, doing something uncommonly useful: documenting free, reliable information and giving it away to hundreds of millions of people every month for free.

But we cannot tend and grow the knowledge commons alone.

In 1833, an English economist wrote about a hypothetical example of how to deplete a common good. If herders sharing common lands allowed their animals to graze, and some always grazed more, the commons could wither.

Although knowledge does not deplete with use, platforms, infrastructure, and peripherals do.

As more minds engage with knowledge, it tends to grow better. Were that only true for infrastructure. More minds brings needed investment in scale—capacity, speed, security, contextual search, and translation. We must continually invest in the renewal of our platforms, infrastructure, and peripherals. And we must grow. It is not enough to maintain the status quo. We must reach beyond our half a billion users. We must reach for seven.

We will need more partners; fewer free riders.

If you run an organization in a knowledge intensive industry—pharma, consulting, tech, materials, energy—it’s very likely your people are daily users of Wikipedia for professional purposes. And if information is central to your core business—search, data, meta-data—then the value of the commons is directly calculable to you.

All that knowledge, all in one place, in 288 languages so that your people don’t have to spend their precious time hunting across the fragmented knowledge landscape… looking here and there. That is, if they can even find what they are looking for and make it over the paywall. The opportunity cost to you is not priceless, but it is significant.

Let's make it mutual.

Whether you employ many workers who use Wikipedia casually, run organizations in knowledge-intensive industries, or information is core to your business model, The Wikimedia Foundation should be honored to have a conversation about how we might all tend and grow the commons together. It will be better for all of us.

A rough heuristic to start.

Is it the number of knowledge workers at a rough percentage of time? Is it an annual fee to contribute to the health of the commons? There's room to experiment. Recent years have basically made the argument for us: we all benefit by tending the commons.

Annaproject (talk) 17:51, 11 August 2018 (UTC)

Generally, I'm in agreement with this proposal. I generally run into Facebook and Google researchers who are interested in building collaborations around information integrity. They have huge indexes and our contributors know how to do crowd review effectively. I think these huge indexes could be very beneficial for hoax detection, copyright violation detection, and various types of article-writing support. But one of the key things that holds us back is the corporate wall of secrecy. If I'm going to collaborate with these researchers, we need to either break through huge bands or red tape or we need to work exclusively with public data -- which defeats the purpose of collaborating with the big index folks. A partnership in this area could help us break through the red tape and really start sharing resources. --EpochFail (talk) 10:55, 16 August 2018 (UTC)

To be clear, I would not personally advocate for the sharing of data in the first few years, though there may be some exception that could make sense. I would begin sharing talent and resource and be engaged in conversation about common problems we could work to solve together. What are the most important problems of the information commons? We don't want to solve just any problems, we want to solve the problems that will make a fundamental difference. After a number of years, it may be clear where and when it is safe to share data, but not to start. Annaproject (talk) 18:48, 4 November 2018 (UTC)

It seems as though no one is monitoring this page. Annaproject (talk) 01:05, 30 August 2018 (UTC)

@Annaproject: I am, but what you wrote is addressed to the Group members, and they are just starting their work. The initial stage should take no more than several weeks. Please be patient, and thanks for understanding. SGrabarczuk (WMF) (talk) 20:01, 31 August 2018 (UTC)