The Potential for AI to Restore Local Community Connectedness, The Bedrock of a Healthy Democracy
by Sarah Friar and Laura Bisesto (1)
Digitally-mediated engagement serves as a foundation for larger societal goals, with AI-enabled platforms accelerating the creation, curation, and consumption of information, reshaping public discourse. The authors emphasize the importance of fostering connections among neighbors to strengthen community ties and address local issues.
AI is poised to profoundly influence democracy, particularly in how local communities connect and engage. The advent of AI-enabled platforms accelerates the creation, curation, and consumption of information, reshaping the landscape of public discourse. These platforms integrate text, video, audio, and interactive elements, personalizing content based on user behavior and preferences, thus enhancing reach and engagement in real time. This increased personalization and dynamic content delivery can potentially revolutionize civic participation, making government processes more transparent and accessible, improving civic literacy, and enabling more responsive governance.
However, AI also brings significant challenges. The amplification of echo chambers, spread of misinformation, and manipulation of public opinion through algorithmic bias pose threats to the integrity of democratic processes. As AI accelerates binary thinking and bot-driven discourse, it may deepen societal divides and erode trust in public institutions. Thus, while AI offers the potential to expand democratic engagement, it also necessitates careful governance, accountability, and ethical standards to mitigate its risks.
A digital divide: Eroding neighbor connections despite a longing for local relationships
A majority of Americans (2) don’t know most or even any of their neighbors. This is not just an American problem—it is a global phenomenon. And yet it doesn’t have to be this way. We have found that change can begin with simple acts of kindness, that gathering together in our neighborhoods can have a profound effect on our lives, and that a large swath of people want to ground themselves in local connections. During a trip to Atlanta, at the beginning of my tenure as CEO of Nextdoor, (3) I pulled up a chair with a dozen people who called the city home and who had connected on social media: neighbors, community leaders, public officials, and local entrepreneurs. Our group came together to discuss a common goal: They had all been working on solutions— individually up until that point—to help neighbors in need of food security. Our conversation was productive. They left with more resources than they came in with, and they left knowing a bit more about neighbors they could rely on.
That table in Atlanta was one of many where I’ve witnessed the real-life impact of online-to-offline connections. I’ve been to neighborhood fiestas in Spain, a local business-owner meetup in Denver, a town hall meeting in Sydney, a community group in Canada, and a neighborhood gathering in my native capital city of Belfast, Northern Ireland. I’ve watched hands, both literally and figuratively, reach across tables, and often across divides, to share problems, ideas, and support. Taking part in neighborhood conversations reinforces what we know to be true: what we have in common is greater than what separates us.
We live in an incredible age where technology has linked billions of people, offering instant access to vast information and diverse opinions. The primary interface for the majority of these connections is through social media platforms. Per DataReportal, (4) there are now 5.04 billion active social media user “identities” in the world, or over 60% of the world’s population, and while this figure may not represent unique individuals, it’s still a representative figure for overall social media use. Social media use also serves as a significant portion of time spent online. In fact, social media users who are 16 to 64 years old spend, on average, roughly two and a half hours online on social media per day. This is almost 16% of their waking hours, assuming around eight hours of sleep per night, and almost double the time they spend eating or drinking. Social media is how we digest news, sports, celebrity gossip, friends and family happenings, and myriad other sources and generation of real-time debate and discussion. As the Economist notes, “social media are the crucible of online debate.” (5) Yet despite these global connections, division and discord feels at a fever pitch, with long-standing assumptions about our social fabric, universal values, and even the nature of truth and trust becoming frayed. Why?
“And as we become increasingly digitally connected, technology increasingly groups us with like-minded people to create our own personal echo chambers. The less we prioritize the intentional building of relationships with people whose worldviews are different from our own, the more we seem to experience tension and a lack of empathy toward people with divergent views and toward the world around us as a whole.”
In so many ways, these connections and this broad access are helpful. But the more we use social media, the more social media personalizes our experiences, naturally curating and organizing the information we consume—in many cases, reinforcing assumptions. And as we become increasingly digitally connected, technology increasingly groups us with like-minded people to create our own personal echo chambers. The less we prioritize the intentional building of relationships with people whose worldviews are different from our own, the more we seem to experience tension and a lack of empathy toward people with divergent views and toward the world around us as a whole.
Enter AI: The advent of AI-enabled platforms further complicates this picture. The last several decades of social media development are likely, in hindsight, to be seen as relatively slow, mostly relying on user-driven sharing. Most platforms started heavily text-dependent and have only recently lightly integrated audio and video. The reach has been significant, as noted above, but limited by network effects and the gradual adoption of technologies. These traditional social media platforms have enabled broader participation in public discourse, allowing voices from varied backgrounds to be heard. However, the lack of sophisticated tools for verifying information has often led to the spread of dis- and misinformation. Finally, engagement to date is still relatively manual and requires significant user initiative.
Now with AI, the creation, curation, and consumption of information is accelerating exponentially. AI-enabled platforms are already characterized as being highly dynamic, with the ability to aid in content creation and seamless integration of text, video, audio, and interactive elements. They enhance the speed at which information is disseminated, amplifying personalization based on behavior, preferences, and social graphs, thus greatly expanding reach and engagement in real time.
The possible benefits of using this technology to increase civic participation could have a significant impact on how local government—and, by extension, democracy—works for the people. AI can increase transparency around, and accessibility to, government processes and services, and allow governments to tailor communications to individual needs and preferences, potentially improving civic literacy and participation. At the same time, it can automate and scale constituent engagement with government, making government and organizations more responsive. However, concerns still remain about echo chambers, dis- and misinformation, and the manipulation of public opinion through algorithmic bias.
Lastly, AI complicates the questions of governance, accountability, and ethical standards in digital platforms. The technology has the potential to even further accelerate binary thinking, bot-driven posting at massive scale, and far less ability to ever hear, let alone trust, the other side of an argument. And this creates a very real issue for communities. While we view AI as having a major impact on our digital lives, the way digital impinges on real life may mean that AI is actually the biggest single influencer of our real-life experience, in particular, how community connectedness is established and nurtured. If we view this form of connection as the bedrock of a healthy democracy, then we can see why AI must be considered a major factor in how democracy evolves from here, as well.
“While we view AI as having a major impact on our digital lives, the way digital impinges on real life may mean that AI is actually the biggest single influencer of our real-life experience, in particular, how community connectedness is established and nurtured.”
The economic and social costs of ignoring the demise of neighborhood connectivity are extremely high, but our data shows that these antisocial trends can be reversed.
Dr. Robert Putnam’s seminal work, Bowling Alone, published over 20 years ago, painted a stark picture of America: declining civic engagement, shrinking social circles, and the erosion of communities and democracy writ large. Putnam argued that technology played a key role in this, pushing us from neighborhood stoops to virtual echo chambers and, ultimately, an epidemic of loneliness. (6)
And now, as it is widely known, the epidemic of loneliness has far-reaching societal and economic impacts. A study (7) conducted by Nextdoor, in partnership with the Observatory of Unwanted Loneliness (created by the ONCE Foundation), (8) observed equally stark findings around the economic costs of loneliness. The study, centered in Spain, found that unwanted loneliness negatively impacts individuals’ quality of life and imposes considerable economic costs on both the national healthcare system and productivity. Specifically:
Unwanted loneliness in Spain creates an economic burden estimated at 14 billion euros per year, representing 1.17% of Spain’s Gross Domestic Product (GDP) in 2021. The costs are mainly attributed to increased use of healthcare services, costing over 5.6 billion euros per year, and drug consumption, costing nearly 496 million euros.
Productivity losses related to unwanted loneliness amount to over 8 billion euros annually, roughly 0.67% of the GDP.
Unwanted loneliness results in a decrease in quality of life, equating to a loss of over 1 million Quality Adjusted Life Years (QALYs), and nearly 18,000 QALYs lost annually due to premature deaths. This equates to a 2.79% reduction in healthy years of life for the Spanish adult population.
However, there are ways to reverse course and meaningfully reduce loneliness around the world. Dr. Kelli Harding, physician and author of The Rabbit Effect, has researched how neighborhoods and communities affect individuals’ well-being and affirms that the most important thing we can do for our health is to foster positive relationships. Dr. Harding shares, “A review of 148 studies with 308,849 participants found that active social circles increased a person’s likelihood of survival by 50% regardless of age, gender, or other medical problems.” (9)
With this in mind, in 2020, Nextdoor embarked on an exploration to see if we could put a brake on the trends driving down neighborly connections. Was an antidote possible? We conducted a global scientific study (10) in conjunction with key academic leaders, including Dr. Julianne Holt-Lunstad, a prominent psychologist and researcher who served as the lead science editor on the U.S. Surgeon General’s Advisory on the Healing Effects of Social Connection and Community, titled Our Epidemic of Loneliness and Isolation. (11) The study examined how meaningful connections and small acts of kindness impact feelings of loneliness, quality of life, and well-being.
The results found that knowing as few as six neighbors reduces the likelihood of feeling lonely and is linked to lower depression, social anxiety, and financial concerns related to COVID-19. Additionally, performing small acts of kindness for neighbors, such as checking in on a neighbor, providing information, or bringing in their trash bins, also reduces the likelihood of feeling lonely. While 1 in 10 participants said they experienced feelings of loneliness at the start of the study, the number was reduced to 1 in 20 after the study was completed. This was clearly a very good start.
Cultivating more casual, low-stakes relationships, like those with your neighbors, is often the beginning of something bigger. A chat at the local coffee shop or in line at the grocers, sharing what you did this weekend with a parent at school pick-up or on a walk with a dog—these are all known as “weak ties,” as described by sociologist Mark Granovetter, (12) though their effects are anything but weak. These everyday moments of connection can help us feel less lonely and more empathetic; they are trusted sources for services, recommendations, and even jobs. They can also guide us in forming political viewpoints or create spaces to debate contentious issues at a local level. In particular, they can act as a way of drawing us out of our echo chambers and into the middle zone, where more constructive alternatives can be found. And stronger neighborhoods have an outsized impact in other areas. In well-connected neighborhoods, fewer lives are lost in tragedies, and communities recover faster from disasters. (13)
What all of this research suggests is that we don’t have a choice—sitting back and allowing technology to ride roughshod over communities and local connection, diminishing their spheres of influence, is too high a price to pay. As David Brooks has said, “the neighborhood is the unit of change”; thus, it is critical to unlock the power of AI to help communities build back their strength and resilience, one neighborly interaction at a time.
AI-enabled platforms can support the creation and sustenance of meaningful, place-based interactions and networks of support: a next leap forward in community building.
For over five years, we had the privilege of helping build Nextdoor’s unique local dataset, which consists primarily of user-generated information from neighborhoods. Data includes real-time insights about local events, safety reports, recommendations, service provider reviews, and discussions happening in a specific locality. It can also include demographic data, and interests and behaviors of the residents in those neighborhoods.
Many platforms, including Nextdoor, have harnessed AI and machine learning (ML) for years to recommend timely and relevant content to users. While many platforms get lost in the vast expanse of global data, Nextdoor delves deep into hyperlocal insights. Coupled with neighbors’ unique insights, Nextdoor’s local dataset can create the ability to tap into the heartbeat of communities, ensuring not just awareness, but an ingrained capacity to help communities thrive. With this data in place, AI can potentially not only create and curate neighborly interactions but also activate communities.
So what could this roadmap for enabling thriving communities entail? From a technical perspective, we have already seen a number of AI-enabled solutions developed, from platform design, to generative AI for more engaging content creation, to nudges for kinder language, to data that leads to even better community engagement outcomes.
AI can add a more dynamic, sophisticated approach to experimenting and iterating on platform architectures, which research suggests can shape a more prosocial outcome.
Social media platforms can be designed to encourage more civil conversations among neighbors, promoting individual and community well-being. A study conducted in collaboration with Yale Law School prompted neighbors on Nextdoor to move their popular posts from their neighborhood-wide feed to new groups dedicated to discussion of that specific topic, and included an experiment that randomized the announcement of platform safety guidelines to members who join these groups. (14) Published in October 2022, the results led to content that was less likely to be reported and more likely to be virtuous in tone, which indicates that platform architecture can be used to positively shape the civility of conversations.
Further, to enhance neighborhood vitality, generative AI can be used to foster constructive conversations by suggesting ways to rephrase potentially offensive comments with more prosocial revisions.
Nextdoor’s Kindness Reminder, a feature designed to detect content that may be uncivil or harmful and then encourage the author to edit their content before they publish, has consistently yielded positive outcomes over the years, with neighbors who encountered the reminder editing or withholding their content 35% of the time in 2021 and 36% of the time in 2022 and 2023. Suggesting a more constructive generative-AI rephrase to content that triggered the Kindness Reminder has also had promising results. Of those who encountered the generative AI-revised text, 26% adopted the suggestion and published more constructive content. Exposure to the Kindness Reminder and the AI Kindness Reminder reduced the creation of uncivil and harmful content by 15%. (15) Hence, AI, when used to promote the composing, curation, and consumption of content, is an important lever to be considered when seeking to promote civility and other prosocial behaviors.
When it comes to generative AI, the opportunities to strengthen human connection and help neighbors bring their best selves to the community are real.
One example from Nextdoor is the integration of generative AI into the content creation flow to help neighbors create engaging and constructive conversations. Built in partnership with OpenAI and leveraging ChatGPT, the Nextdoor “Assistant” feature enhances how neighbors communicate with one another by helping them write posts that are more likely to drive positive community engagement. We found that more than 70% of neighbors who opted to use generative AI to edit their posts accepted the generative AI suggestion, and that small businesses used the generative AI Assistant at twice the rate of other neighbors.
Notifications, nudges, and other communication channels can force multiply engagement.
One of the ways to drive user engagement in online communities is through emails and text notifications, where users are sent a single post from the neighborhood that a user might be interested in and want to engage with, therefore activating further community outreach. By fine-tuning ChatGPT to generate informative and interesting email subject lines, more emails are opened and clicked on, eventually driving more community engagements.
Businesses can also see a positive outcome from the use of AI to better serve the local communities in which they operate.
A neighborhood is much more than just the people who live there. A healthy neighborhood comprises local businesses, from the coffee shop on the corner to globally recognized brands such as Walmart and McDonalds, as well as local schools, churches, nonprofits, and public agencies; the list is long and diverse neighborhood by neighborhood. Local businesses are the lifeblood of communities, and this is another fertile area to leverage AI to make communities stronger.
“We found that more than 70% of neighbors who opted to use generative AI to edit their posts accepted the generative AI suggestion, and that small businesses used the generative AI Assistant at twice the rate of other neighbors.”
Though many local entrepreneurs don’t have the luxury of time or expertise for in-depth marketing, they can employ generative AI to turn them into successful marketers using the most modern tools.
AI can capture and summarize recommendations from neighbors, providing richer, more localized insights than mainstream review platforms. Building on these local insights, AI can turn those reviews into an efficient way to create an ad on the platform. AI has the ability to make it simple and fast for that business to confidently present itself and to focus not on creating marketing materials, but rather doing what they need to do: make their beautiful products or perform their incredible services, to entice more customers and generate more revenue.
Platforms can increasingly leverage natural data tagging at massive scale and minimal cost to enhance local government and local engagement.
For example, in social media and even on search platforms, users are already organically tagging data when they highlight a local event, sell an item, or flag content for moderation. This organically tagged data serves as a valuable training set for any specialized large language model. These specialized models can integrate real-time, location-specific knowledge unique to each community, holding immense potential for creating and sustaining robust, place-based interaction networks and support systems. For example:
1. Hyper-localized Service Recommendations: The specialized AI model could analyze community interactions and preferences to offer personalized recommendations for local services and vendors. For instance, if many users in a region discuss plumbing issues during a cold snap, the AI could proactively recommend trusted local plumbers who are highly rated by the community.
2. Event Awareness and Participation: By tagging data related to local events, the AI can help boost awareness and participation. The AI could automatically inform users about upcoming events tailored to their specific interests and previous engagements, encouraging a deeper community connection.
3. Localized Emergency Alerts and Resources: In emergencies, the AI could act as a real-time information hub, analyzing tagged data to provide updates, safety tips, and resource locations (like water distribution centers during a flood). This would make it a valuable tool for community resilience.
4. Community Governance and Planning: The AI could analyze datatagging trends to help local governments and organizations understand community needs and concerns, facilitating more informed decisionmaking. For instance, frequent tags about park safety could prompt a local government response to improve lighting or security.
5. Support Network Formation: The AI could identify and connect individuals facing similar challenges or life situations based on their interactions and data tags, fostering peer-support networks within the community. For example, new parents or people caring for aging relatives could find and support each other through coordinated local groups.
6. Local Economic Development: The AI could assist in identifying economic opportunities by analyzing community interest and needs tags, supporting local entrepreneurs with data-driven insights into potential markets and services in demand.
We are just at the beginning here, but these technical examples demonstrate what we are already seeing: the potential for AI’s positive impact on community building. In a world where the conversation about AI leans toward binary thinking—world-ending or the solution to everything—these use cases for AI are evidence that a more nuanced approach combining humans with machines, EQ and IQ, helps everyone in a neighborhood get things done, while at the same time directing those online toward beneficial and more community-minded actions. Thus, online action becomes a strong lead for offline outcomes, like the ones witnessed around the neighborhood table in Atlanta.
However, risks are myriad, and we must be proactive to ensure digital equality and preserve what makes local extraordinary. Any future AI roadmap for community building should draw in cross-sector partners and promote a thoughtful build-out of government investment, and regulatory oversight will be required.
The discussion above paints a fairly optimistic picture of how technology with AI as an accelerant really could be the answer to our lack of community connectedness. However, there are also many potential pitfalls that developers of AI need to both be aware of and actively respond to.
“While the urgency of closing the digital divide is already a well-known global issue, the need to promote equitable access to AI technologies has just upped the ante.”
First, we must consider the ethical implications. With the increase in the prevalence of AI in everything that we do, we must guard against a new “digital divide,” ensuring that everyone, irrespective of background, gender, or ethnicity, has access to these advancements. The pace of generative AI adoption is staggering, with ChatGPT forecast to have an adoption curve even steeper than smartphones. Like every digital breakthrough before it—telephones, TV, Internet 1.0, the rollout of broadband, the mobile revolution—the opportunities enabled by AI leave those without internet access further behind just by virtue of where they are born. While the urgency of closing the digital divide is already a well-known global issue, the need to promote equitable access to AI technologies has just upped the ante. Billions of people around the world could be excluded from the benefits of this technological revolution if AI is not thought of as any other necessity—access to food, water, etc. Global investment by governments and NGOs, particularly in more rural areas and the Majority World, is urgently needed to narrow the current digital divide and ensure these communities are not left out of access to this transformative technology.
Further, when building on a local dataset, it’s important to be mindful of and actively celebrate the rich tapestry of local dialects and expressions without imposing a homogenized linguistic model. This is essential, because localism isn’t just about geographic data; it’s about embracing the nuances, languages, and cultures of each community. As part of the future roadmap, it’s imperative that technologists not just prioritize training on datasets from diverse sources, but also continue to work with cross-sector partners including civil rights leaders, social scientists, and behavioral scientists to discuss issues such as fine-tuning generative AI models to uphold unique local speech patterns. It’s also important to create, where possible, feedback mechanisms for users to report instances where AI-generated content does not perform as intended.
“As part of the future roadmap, it’s imperative that technologists not just prioritize training on datasets from diverse sources, but also continue to work with cross-sector partners including civil rights leaders, social scientists, and behavioral scientists to discuss issues such as fine-tuning generative AI models to uphold unique local speech patterns.”
Finally, being able to truly understand human- versus machine-created content in a local context is also important. In some cases, a machine’s ability to synthesize and quickly return an answer to a question could significantly outperform a single human sifting through many websites. Think of a question like “where best to have my five-year-old’s birthday party this weekend?” It may not matter whether a human provides a response here, or a machine. However, replies driven by AI to questions about local candidates in an election, or what to do about societal issues such as the unhoused, might have the capacity to change neighborhood opinions en masse. Transparency is critical. Therefore, it is critical that there is transparency around when users are interacting with a human versus a machine, as well as continued investments in, and adoption of, content provenance methods to strengthen the integrity of digital content. We note that key stakeholders are already creating and joining organizations such as the Coalition for Content Provenance and Authenticity (C2PA). C2PA is a widely used standard for digital content certification, developed and adopted by a wide range of actors including software companies, camera manufacturers, and online platforms. We also recognize that one size doesn’t fit all when it comes to AI ethics, and there’s an equally important role for governments to play in mitigating risks of harm by AI—from national security to individuals rights. Governments should engage in an informed decision-making process that is both scientifically rigorous and inclusive, to develop robust standards and guidelines that have the greatest potential to promote innovation while safeguarding against potential harms.
Looking to the future
As we look forward, AI can clearly be a powerful tool for building strong local communities by fostering meaningful connections, but it also has the potential to further divide if not carefully designed and implemented. There is no straightforward answer and, as with any complex system, many of the consequences will not be immediately evident, months or even years after deployment. What we can control is the approach and collaboration needed to ensure the right guardrails are in place. As online platforms, including social networks, consider ways to leverage AI, we urge them to consider the promise it holds to build communities, one neighbor at a time, no matter how small or mundane a connection may seem. As they build, we urge them to engage a collaboration of diverse, cross-disciplinary, intergenerational stakeholders to nurture equity and inclusion. Most importantly, we must all hold ourselves accountable to future generations to preserve what makes our communities resilient and uniquely local.
While community is the bedrock upon which a thriving democracy can flourish, community in and of itself is not sufficient for a democracy. In this transformative era, let’s also commit to harnessing AI as a force in support of democracy. Let’s develop these new technologies with a strong point of view on how AI can be used to amplify diverse voices while bridging divides, and build a future where technology strengthens our bonds and enhances our civic life. The time to act is now—let’s ensure that AI serves as a cornerstone for a vibrant, inclusive, and resilient democracy.
Footnotes
(1) The authors would like to thank Kiran Prasad, Julianne Holt-Lunstad, Kelli J. Harding, Jared A. Crooks, Antonia Gray, Craig Lisowski, Tiger Zhang, Qi He, and Jaewon Yang.
(2) A majority of Americans (57%) say they know only some of their neighbors; far fewer (26%) say they know most of them. Americans ages 65 and older are more likely than those ages 18 to 29 to say they know most of their neighbors (34% vs. 20%). In contrast, about a quarter (23%) of adults under 30 don’t know any of their neighbors, compared with just 4% among those 65 and older. Leslie Davis and Kim Parker, “A Half-Century after ‘Mister Rogers’ Debut, 5 Facts about Neighbors in U.S.,” Pew Research Center, August 15, 2019, https://www.pewresearch.org/short-reads/2019/08/15/facts-aboutneighbors- in-u-s/.
(3) Nextdoor Holdings, Inc., is a U.S. public company that operates a hyperlocal social networking service for neighborhoods. The company was founded in 2008 and is based in San Francisco, California. Nextdoor launched in the United States in October 2011, and is available in 11 countries as of May 2024, serving close to 100 million neighbors across 335,000 neighborhoods.
(4) Simon Kemp, “The Time We Spend on Social Media,” DataReportal, January 31, 2024, https://datareportal.com/reports/digital-2024-deep-dive-the-time-we-spend-onsocial- media.
(5) “The End of the Social Network,” Economist, February 1, 2024, https://www.economist. com/leaders/2024/02/01/the-end-of-the-social-network.
(6) Robert Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon & Schuster, 2000).
(7) Bruno Casal Rodríguez, Berta Rivera Castiñeira, and Eva Rodríguez-Míguez, The Cost of Loneliness in Spain (San Francisco: Nextdoor and Madrid: Observatorio Estatal de la Soledad No Deseada, 2023), https://www.soledades.es/sites/default/files/contenidos/ Executive%20summary_The%20cost%20of%20loneliness%20in%20Spain.pdf.
(8) “Observatorio Estatal de la Soledad No Deseada” (“Observatory of Unwanted Loneliness”), accessed July 3, 2024, https://www.soledades.es; “Foundación ONCE” (ONCE Foundation),” accessed July 3, 2024, https://www.fundaciononce.es/es.
(9) Kelli Harding, The Rabbit Effect: Live Longer, Happier, and Healthier with the Groundbreaking Science of Kindness (New York: Atria Books, 2019).
(10) Michelle H. Lim, Pamela Qualter, Alexandra Hennessey, Ben J. Smith, Taylah Argent, and Julianne Holt-Lunstad, “A Randomized Controlled Trial of the Nextdoor Kind Challenge: A Study Protocol,” BMC Public Health 21, no. 1510 (August 5, 2021), doi:10.1186/s12889-021-11489-y.
(11) The U.S. Surgeon General’s Advisory on the Healing Effects of Social Connection and Community: Our Epidemic of Loneliness and Isolation (Washington, DC: Office of the U.S. Surgeon General, 2023), https://www.hhs.gov/sites/default/files/surgeon-generalsocial- connection-advisory.pdf.
(12) Mark S. Granovetter, “The Strength of Weak Ties,” American Journal of Sociology 78, no. 6 (May 1973): 1360-1380, https://www.jstor.org/stable/2776392.
(13) A 2021 study by Courtney Page-Tan published in the International Journal of Disaster Risk Reduction analyzed how social media use impacted recovery efforts in the weeks following 2017’s Hurricane Harvey. Communities that were more active on Nextdoor were more likely to recover faster than communities that were not active on Nextdoor. Courtney Page-Tan, “An Analysis of Social Media Use and Neighbor-Assisted Debris Removal in Houston Following Hurricane Harvey, International Journal of Disaster Risk Reduction 63, 102450 (September 2021), https://www.sciencedirect.com/science/ article/abs/pii/S2212420921004118.
(14) Team Nextdoor, “Nextdoor Vitality Efforts Published in New Study in Partnership with Yale Law School,” October 14, 2022, https://blog.nextdoor.com/2022/10/14/nextdoorvitality- efforts-published-in-new-study-in-partnership-with-yale-law-school/.
(15) Team Nextdoor, Transparency Report 2023 (Nextdoor, February 2024), https://about. nextdoor.com/wp-content/uploads/2024/02/Nextdoor-Transparency-Report- 2024-2023-Reflection.pdf.