By Ariadna Vargas and Jeremy Barofsky

Memphis, Tennessee, is often the subject of negative media attention, particularly when it comes to crime and poverty. However, after working for months with local non-profit organizations as part of our Narrative Change work, we were able to confirm what residents of the city have always known—that there is so much more to Memphis than the media headlines suggest. Jennifer, a young, proud Memphian who lives in the heart of the city that she describes as “lively, diverse, and charming” put it best: “Memphis is just something you have to see to believe. There are too many negative news stories that don’t accurately portray what it’s like to spend a day in the area.”

While Memphis is undoubtedly a city bursting with culture, diversity, and liveliness, it also faces a range of challenges, including racial equity, crime, and poverty. Unfortunately, solutions to these problems are often imposed by outsiders, neglecting to consider the unique challenges of each community and failing to build upon their distinctive strengths. It raises the question: In light of limited resources, how can local governments truly connect with, understand, and make choices that reflect the diverse viewpoints, preferences, and needs of their constituents?

To really combat the complex drivers that create and sustain poverty, it’s crucial to learn directly from the real experts—those experiencing it first-hand. With this in mind, we partnered with six local grassroots organizations that are deeply committed to supporting Memphians and promoting community engagement. We also connected with an extensive informal network of non-profits, academics, and advocacy groups that share our goal of building effective data-driven solutions to create a brighter future for our communities. We recognized that traditional survey methods wouldn’t be enough if we wanted to capture the depth and nuance of people’s experiences. That is why we developed an open-ended survey that encouraged people to express their genuine thoughts and feelings in their own words, and to fully elaborate on their experiences and perspectives. By using this flexible format, we were able to gain a more nuanced understanding of people’s viewpoints and uncover insights that might have been missed with a more structured approach. Additionally, by spreading the word via our networks and partners, we were able to gather a remarkable range of insights from 418 individuals of various backgrounds who live in 52 different zip codes within Shelby County, Tennessee.

Our subsequent research centered on two questions from the survey, which delved into the most commonly discussed social issues in Memphis, and the necessary changes required to improve the city’s image in the eyes of others. To analyze the responses, we used a method called natural language processing (NLP), which is a subfield of artificial intelligence (AI) that applies algorithms to understand and decipher human language. 

The results from this exercise validated findings from Community Circles, where we engaged in discussions with local residents to understand their perspectives on their own communities. These conversations revealed that crime, violence, and poverty are by far the biggest issues on residents’ minds. The chart below shows the main issues mentioned by survey respondents, as well as the top keywords associated with each issue.

Despite being critical of their city, Memphians also have a strong sense of pride in their hometown, as reflected in the comments of one study participant, Jana, who described the city as “the people you bump into, weird, over-the-top friendly, nuts … we are one big, dysfunctional family, full of love and energy that you must experience to understand!” Her views of Memphis also summarize the perceptions of most participants by suggesting that to change the negative view of outsiders, they first need to “change some of the actual problems, like poverty, crime, education.” Additionally, her comments suggested locals should shift their mindset and attitudes toward their hometown, and get involved in their city’s progress to observe “the good things happening all around us.” Finally, she suggested that Memphians take responsibility for the narrative they project to others by avoiding “perpetuating negative stereotypes, and being mindful of the impact of their self-deprecating jokes and complaints on Memphis’s image.” 

The following chart shows the solutions that were most frequently mentioned by survey respondents to improve other people’s perceptions about their city. In addition to addressing issues related to crime and poverty, the idea of “experiencing true Memphis” was also frequently suggested as a way of improving the city’s reputation.

In summary, our study showed that using natural language processing was an effective way to efficiently analyze large amounts of text-based data. From that analysis, we could extract valuable insights into the concerns and sentiments of community members that can inform more equitable and effective solutions. Ultimately, these are what will make a real difference in people’s lives.

In this context, the power of large language models such as the recently launched GPT-4 can revolutionize the way local governments gather and analyze opinions from their residents. This approach can be a game changer for allocating resources, creating new programs, and investigating other areas for social impact. As these AI tools become more accessible and prevalent, local governments can become more responsive to their citizens’ needs and preferences, promoting programs and policies that enhance resident autonomy and dignity. 

It’s crucial, however, to use these models responsibly and be mindful of the potential harms they can cause if misused. By using these tools cautiously, local governments can take significant steps toward improving the lives of their citizens and implementing equitable policies that make a positive impact on their lives.