A translatable version of this post can be found on Meta-wiki
The Wikimedia movement has created one of the top websites in the world. Wikipedia, the encyclopedia that anyone can edit, is read more than 15 billion times a month. Wikipedia and its sister projects represent the backbone of an entire ecosystem of free knowledge – a global movement of volunteer contributors that, over the past two decades, have captured the largest collection of knowledge in human history. As we look to the next two decades, we must continue to ground our work in the realities of a rapidly-changing world in order for this movement to continue creating the world’s knowledge ecosystem.
Two weeks ago, the new Wikimedia Foundation CEO Maryana Iskander shared her takeaways from a two-month-long listening tour speaking to Wikimedia contributors, staff, and partners. Her first question centered around the need to understand how the movement can stay relevant in the face of new external trends, in order to ensure our principles of open knowledge creation survive into the future. As a movement, we must ask ourselves, “What does the world need from us now?”
To begin, a group of Foundation leaders set out to identify relevant trends that impact the information ecosystem, both now and quite likely into the future. These are trends that we should expect to accelerate in the years to come because they relate to key changes in how people access, interact with, and share knowledge.
They can present both opportunities and challenges for our movement.
Search has fundamentally changed
Both what users are searching for, and how they are searching for it have fundamentally changed over the past several years. These changes in information seeking behavior will likely influence how new readers come to Wikipedia and other free-knowledge projects.
What people are searching for is changing: Users increasingly expect answers to their search queries in rich content (e.g., image, video, and audio formats).
Rich content-first platforms like TikTok, Instagram and YouTube are being used as entry points for information-seekers (that is, as search engines) and have amassed huge audiences. In 2021, for example, TikTok surpassed Google in traffic, becoming the world’s most visited web domain.
Taking advantage of this trend, online platforms are actively investing in creating rich, sticky content, ensuring searchers come to – and stay on – their platforms. This is a change from traditional search engines, which point the user to third party content to answer their queries. For example, TikTok is attracting educators and subject-matter experts to create educational content through new features and campaigns, e.g. Jumps and #learnontiktok. (See more about this in the next section.)
In an effort to not be left behind, traditional search engines are investing in visual-first search experiences as well. Google is increasingly serving both text and imagery directly on the knowledge panel, hoping users come to Google and stay there when they find what they need.
How people are searching is changing: Voice is becoming the dominant way that people search the web.
Today, 30% of all searches globally are being done using a device without a screen. In 2018, nearly a third of the global population was using voice search on mobile, and by 2020, that number had reached more than 50%. In the U.S., household penetration for smart speakers is predicted to rise to 55% this year.
As a platform that relies on most of its traffic from traditional text search, what do these shifts in search behavior mean for our current and future work?
Meeting the global demand for content
The number of Internet users has doubled in the last decade, with most of this growth driven by the newly-connected in Asia and Africa. But more than half of Internet content today is in English, a language spoken by less than 20% of the world’s population. Internet platforms are investing massive resources in creating original, local-language content to remain competitive.
They are doing so in two primary ways:
Increasing investment in content creation and content translation
Users are incentivized to create content on popular platforms (e.g., YouTube, Substack) through monetization channels (sponsorships, subscriptions, ad sales) that can deliver substantial revenue for successful creators. Facebook, Snap, and Twitter have all recently announced plans to provide monetary incentives for users who hit certain content and/or follower thresholds, to continue to keep top creators engaged and recruit new power users to their platforms. In addition, in the last decade, the number of people employed in the translation industry has doubled, largely to meet new consumer technology demand.
Investing in machine translation and Artificial Intelligence (AI) content creation
AI has gotten significantly better at translating content, and even at creating new content. For example, Google has made automatic machine interface translation a core feature of its Chrome browser: users searching for content that does not exist in their language increasingly see machine-translated English website content (including Wikipedia). It has also invested in AI to generate new content (e.g. Google Brain Wikipedia articles/stubs). Improvements in AI have the potential to radically transform both the translation and creation of new content. Existing technology players like Google and newer organizations like OpenAI are competing to build more robust and flexible content-generating AI engines, which may in the near future be capable of sophisticated knowledge synthesis and creation (though it is still unclear when/if these technologies will perform at or near human levels).
How can Wikimedia projects ensure that relevant, local-language knowledge gaps are filled? How do we continue to raise awareness of our community-driven model of content creation that is different to other platforms?
Disinformation and misinformation are on the rise
Disinformation and misinformation are growing, with unreliable sources doubling their share of social media engagement in 2020 as compared to 2019. Disinformation and misinformation are spreading quickly due to technology that allows for it and changes in user behavior, including online reading that is increasingly characterized by scanning, browsing, and non-linear reading.
Technology platforms are taking three primary approaches to tackling the disinformation crisis – many borrowing from Wikimedia processes or relying on Wikimedia for ground-truthing – but with limited success:
Paying human moderators to complement AI algorithms
TikTok, for example, is scaling up the 10,000 content moderators it currently employs. YouTube, Pinterest, and others are similarly scaling their human content moderation capacity.
Leveraging Wikipedia as a fact-checking source
YouTube’s InfoBanners pull information from third-party sources, including Wikipedia, to provide context on disinformation-prone videos. Google and Facebook use Wikipedia hovercards to provide more context on potentially misleading information.
Experimenting with community moderation
Borrowing explicitly from the Wikipedia model, Twitter’s new Birdwatch product asks power users to identify Tweets they believe are misleading and attach public notes to provide informative context.
With misinformation and disinformation threatening the quality of information on Wikimedia projects and on other platforms, what role, if any, should the Wikimedia movement play in addressing disinformation in a wider knowledge ecosystem?
Each of these trends has the potential to shape the role of Wikimedia projects within the larger information ecosystem. In an increasingly complex, interconnected future, we will have to anticipate other influential factors, including the changing landscape of internet access and expansive government regulations that seek to legislate issues from privacy to content moderation to the influence of social media platforms on misinformation. (For example, this recently published blog post highlights the risks of upcoming legislation and the importance of advocating for policies that protect free speech and the free knowledge ecosystem.)
With over two decades of experience already, this movement can continue to make significant and positive contributions in collecting and sharing free knowledge. These and other trends should inform our approaches to implementation as we march towards a bold vision in 2030 of further expanding free knowledge to societies around the world.