About this study
About this study
Why we did this study
This project is a continuation of CNTI’s Defining News Initiative which seeks to understand how journalism is defined today. Technology, social media and AI have significantly influenced how journalism is defined which impacts its role in democracy.
Over the last few decades, the widespread adoption and use of new communications technology have rapidly reshaped global information ecosystems. For journalists and news organizations, these changes have both empowered their work and presented new challenges. These technological and social changes have given the public many more options, disrupting how news outlets have historically interacted with them. Concurrently, governments around the world are increasingly impinging on press freedom and weaponizing the law against journalists, while questions about revenue models and digital content valuation remain unresolved.
Social media platforms have offered new opportunities to meet the public where they are, but social media has also decoupled distribution from news outlets’ function as a clearinghouse for vetted and verified information. It has never been easier for journalists to have meaningful exchanges with their audiences; however, these same channels leave journalists on the receiving end of continuous legal threats and harassment.
Most recently, the convenience and expediency of AI1 tools help individuals and teams produce more content — but these tools are resource-intensive and prone to inaccuracy, and the opacity of their algorithms can leave journalists unsure about how their own work is being repurposed. While these technologies can be harnessed to support the work of journalism, they also raise a host of ethical and legal questions that threaten the field.
Access to information is not just important for its own sake; it makes democracy possible. In 2024, the United Nations outlined Global Principles for Information Integrity in response to growing challenges around misinformation, disinformation and hate speech. These principles highlight the importance of societal trust and resilience; healthy incentives; public empowerment; independent, free and pluralistic media; and transportation and research. Multiple actors, such as technology companies, advertisers, news media and journalists, civil society organizations, states and political actors, and the UN play a key role in upholding the resiliency of the information ecosystem and their good faith efforts are essential to the public’s ability to access information.
Building on the Defining News Initiative, CNTI conducted a survey of journalists to explore
- how they view their industry,
- their (and their organizations’) uses and perceptions of technology,
- their perspectives on government action and cyber security and
- their experiences with online harassment and abuse.
Importantly, respondents to CNTI’s study come from a global mix of journalists that provide an international perspective to how journalists are navigating, understanding and defining their rapidly changing industry.
How we recruited participants
Surveys are a snapshot of what people think at a particular moment in time. These data were collected between October 14, 2024 and November 24, 2024, which means they highlight the perspective of journalists around the world during that time frame.
CNTI partnered with journalism organizations in multiple continents. The questionnaire was crafted with input from the partner organizations who knew the current situations and challenges of their members across various country contexts. These partner organizations also shared the survey with their membership:
- Centre for Journalism Innovation & Development (CJID)
- Frontline Freelance Mexico
- FT Strategies
- Global Forum for Media Development (GFMD)
- Global Investigative Journalism Network (GIJN)
- Internews
- Online News Association (ONA)
- Organización Editorial Mexicana (OEM)
- Society of Freelance Journalists
The survey was also shared through several journalism email groups and Slack spaces, including:
- Institute for Nonprofit News (INN)
- News Nerdery
- Newspack Community
- Shorenstein Center on Media, Politics and Public Policy
Who took the survey
CNTI collected responses from 433 journalists across 63 countries between October 14 and November 24, 2024.2 Because there is no global census of journalists it is not possible to create a representative sample. So, CNTI’s results should be interpreted based on the group of journalists who took part in the survey.
The journalists in our survey hold a range of jobs. They are about equally divided between content specialists (e.g., reporters, photojournalists and bloggers), content managers (e.g., news directors and editors) and people in hybrid or other roles. A strong majority of them — 80% or more — work full-time, in-house at a single news organization and nine out of ten say their audience is located primarily in the same country as them.
The majority of respondents are relatively experienced journalists: 61% have more than 10 years of experience in journalism.
While we received responses from all regions of the globe, some places had more respondents than others. For example, 256 respondents are from just three countries: Nigeria, the U.S. and Mexico. The other 177 respondents are from 60 different countries with the highest number of responses in a single country at 23. We do not provide exact numbers of responses for countries that have less than 25 responses to minimize the risk of participants being identified.
Region | N | % |
---|---|---|
Latin America & Caribbean | 124 | 30% |
Sub-Saharan Africa | 102 | 24% |
Europe & Central Asia | 77 | 18% |
North America | 74 | 18% |
East Asia & Pacific | 25 | 6% |
South Asia | 5 | 1% |
To better understand the results, we grouped countries into two larger categories:3
Broad region: Like many other global research studies, we grouped countries into Global North and Global South, which share many demographic, economic and political similarities. In general, countries in the Global North are wealthier, have older populations and wield disproportionate geopolitical power. While the conceptualization of Global South has received criticism for its melding of diverse countries into one binary group, the term is helpful for understanding similarities in country development, history and overall geopolitical perspective. A little more than a third (38%) of our respondents live in the Global North, and the rest live in the Global South (62%).4
Regime type: In addition to Global North and Global South, we used the Regimes of the World (RoW) measure found in the Varieties of Democracy (V-Dem) data which responds to the question “How can the political regime be classified considering the competitiveness of access to power as well as liberal principles?” Electoral democracies are separated from liberal democracies because they lack at least some aspect of personal liberties, transparent law procedures or judicial and legislative constraints on an executive leader typically found in liberal democracies. Autocracies, in contrast, either have no multiparty elections or do not have elections that are free and fair and are missing aspects of freedom of expression found in democracies. Of our survey respondents, 27% live in an autocratic country, 35% live in a liberal democracy and 39% live in an electoral democracy.5
We include these categories in our analysis of cyber security and safety because regime type is associated with harassment (though not necessarily by government actors), the censorship of journalists and the ability of the rule of law to protect them. Democracies tend to have more robust protections for civil liberties like freedom of expression compared with autocracies — a protection that is central to an informed public.
How we addressed attrition
The survey was lengthy (with an estimated completion time of 20 minutes) and consisted of five distinct sections on different topics. We treated each section as a drop-off point: that is, if a respondent answered at least one question within a section, non-responses were treated as true non-responses. Respondents who answered no questions within a section were not included within the section’s N.
For transparency, topline tables include both percentages of the full survey N (Percent) as well as percentages of the section N (Valid Percent).
How we tested for statistical significance
We analyzed the results using Chi-squared proportion tests to assess differences in responses. We used a standard threshold of p < 0.05 for assessing statistical significance. Differences mentioned in the report text are statistically significant.
How we protected our data
CNTI did not collect any identifiable information that risked the privacy and confidentiality of participants. The data collection was supervised by CNTI staff only. The survey included individual-level information such as gender and the country where one worked and resided. It would be very difficult to identify study participants because CNTI did not collect their personal contact information or contact participants directly and the participants’ personal information was not shared by the partner organizations. The data are securely stored in an encrypted folder which is only authorized to the core research team at CNTI.
The survey was administered using Qualtrics online survey platform. CNTI implemented several protocols to maintain privacy — including removing IP addresses from individual responses — and settings to avoid duplicate responses from a given person. Alpha IRB provided independent ethical review.
How we coded open-ended items
A few items allowed for comment if the respondent selected “other” and four questions on our survey were fully open-ended. Because we were conservative in drawing inferences, we note that themes were present in “at least” the number of responses where we identified them.
We were conservative in large part because of the multilingual nature of our data set. Terms do not necessarily have exact equivalents between languages. Moreover, many responses were very brief (five words or fewer) and the absence of context made interpretation challenging even within a single language.
The survey itself was available in nine languages (Arabic, English, French, Hindi, Portuguese, Russian, Spanish, Swahili and Ukrainian). We received free-text responses in six of these languages (Arabic, English, French, Portuguese, Spanish and Ukrainian), as well as one response in an additional language (German). All responses were translated to English by the same translation company who prepared the questionnaire, with the exception of the German response, which was translated in-house by a researcher working on the project.
Wherever possible, categories were informed by preceding CNTI and external research.
We did not receive enough responses to any individual item for robust tests of inter-coder reliability; instead, we reached consensus through repeated discussion of responses.
Specific analysis methods for each of these four questions are outlined below.
Technology Trends
We asked, “What technology trend or development related to journalism is most on your mind these days?”
After researchers read through the full set of responses, these responses were coded for the presence or absence of four main ideas: artificial intelligence, social media, algorithms and data. These categories were not treated as mutually exclusive. Responses were also coded for their valence (positive, negative, neutral or mixed).
Definitions
We asked, “In just a few words, how do you define journalism?”
We first coded responses in terms of whether they referred to the social role of journalism or professional practices and norms. These themes were not treated as mutually exclusive.
- Social role of journalism: Responses that focus on the impact of journalism on audiences or society fall under this theme, as do responses that highlight journalism’s relationships to ideals like “democracy” and “the public interest.” (Focusing on journalism’s power to educate, inform, or entertain people would all fall into this category — as would focusing on journalism’s power to hold powerful actors accountable or support decision-making.) In general, these responses are about what journalism does.
- Professional practice and norms: Responses coded under this theme focus on the process of journalism and professional norms widely shared among journalists. Examples of processes include fact-checking, publishing stories and obtaining information. Examples of norms include that journalism should be objective, verifiable, truthful and independent. In general, these responses are about how journalism is done.
A small number of responses were too ambiguous to code, and several additional responses focused on downsides of journalism, especially poor pay and safety risks.
Category | Number of responses | % of total |
---|---|---|
Professional practice | 223 | 69% |
Social function | 181 | 56% |
Neither of these (including downsides) | 9 | 3% |
Total | 323 | |
Too brief or unclear to code | 30 |
We analyzed the 181 responses that addressed the social function of journalism using six categories defined by the Journalistic Role Performance project (JRP). Meanwhile, we took a bottom-up approach to the 223 responses that addressed professional practice to identify categories that were emergent in our data, starting from oft-repeated words and phrases, which provided five additional categories.
Rather than treating categories as mutually exclusive, we also coded them independently of one another.
Social Functions
- Interventionist: The JRP defines intervention journalism as “a kind of journalism where the journalist has an explicit voice in the story, and sometimes acts as an advocate for individuals or groups in society.”6 For responses that fall under at least this category, cues included journalists explicitly presenting their point of view; explaining or interpreting causes, meanings, and consequences; or offering calls to action.
- Watchdog: The watchdog role is defined as functioning “to protect the public interest and to hold various elites in power accountable, serving as a ‘fourth estate.’ Journalism performance closer to the watchdog role entails being a custodian of conscience, making visible facts hidden by those in power.” Answers in this category included holding those in power accountable, other references to accountability or uncovering things that are hidden.
- Loyal facilitator: As defined by the JRP, journalists acting as loyal facilitators include two variations. “First, journalists cooperate with those in power, and accept the information they provide as credible … In its second variation, journalists support their nation-state, portraying a positive image of their country, encouraging a sense of belonging, and strengthening national prestige.” Cues for definitions in this category included praising, promoting or defending government actions or policies, presenting positive images of those in power or explicit patriotism.
- Service: According to the JRP, “this role combines the rights and self-interests of the audience, creating a client-professional relationship between the journalist and the public. As an answer to the growing complexity of modernity, this model of role performance provides helpful information, knowledge and advice about goods and services that audiences can apply in their day-to-day lives.” Answers coded under the service category mentioned references to consequences for individuals’ everyday life (not at the societal level), references to tips or advice or specific personal appeals.
- Infotainment: Infotainment is defined as “journalism [that] uses different stylistics, narrative and/or visual discourses in order to entertain and thrill the public … the logic here is to shock the audience’s moral and aesthetic sensibilities. This type of journalism addresses the public as spectators, where the audience’s relaxation and emotional experiences become the center of attention.” Cues for this category included details about individuals’ private lives, sensationalism, references or appeals to emotion or graphically morbid details.
- Civics: The JRP defines the civic role of journalism as a focus on “the connection between journalism, the citizenry and public life. Journalistic performance that reflects these ideas encourages the public to get involved in public debate, and to participate in social, political and cultural life. The space given to sources without social empowerment who demand recognition or reinstatement of a right is an important aspect of this role. This role does not assume that it is the journalist who can create an improved community via their own resources; rather the emphasis is on supporting the citizens’ efforts to do so.” Answers categorized in this section included mentions of getting involved in public debate, change at the societal level, local impact, contextual background for citizen decisions and supporting citizen movements.
Professional Practice
Responses focused on professional practice were coded into at least one of the following five categories. The first three categories focus on processes that journalists follow, while the latter two emphasize professional norms.
- Journalism requires newsgathering. One core journalistic process is conducting research and investigation to learn about things that are happening.
- Journalism requires verification. A second core journalistic process is verifying and assessing the truth of that information before sharing it.
- Journalism requires distribution. A third core journalistic process is distributing or disseminating those stories. These responses typically focused on large-scale broadcast communication, which was almost always unidirectional.
- Journalism is factual. A core norm is that journalism is based on facts that are reliable and truthful.
- Journalism is objective. A core norm is that journalism is objective, fair and independent of political or economic influence.
Journalists’ Traits
We asked, “What top three traits or characteristics do you most associate with the job of a journalist?”
The research team read these responses and discussed possible groupings. Because of the brevity of most responses — most of them were just one word long — we identified clusters associated with particular words and phrases.
Here we provide the major cues to each cluster.
Journalists rely on verifiable facts
- truth, truthful, fact, factual
- credible, credibility, accurate, accuracy, reliable, reliability
- trust, trustworthy
- precise, precision
- verify, verification, verified
Journalists have clear ethical standards
- ethics, ethical
- professional, standards
- independent, independence, incorruptible
- honest, honesty, transparency, sincere, sincerity
- accountable, accountability, public interest, commitment to [community/public], watchdog
Journalists are objective
- objective, objectivity, fair, fairness, balance, balanced, neutral, impartial, impartiality, unbiased
Journalists are analytical
- research, investigation, analysis, investigate, analytical, logic
- skeptical, cynical, critical
- gather, dig
- context
Journalists have “grit” and work hard
- grit
- persistent, persistence, perseverance, tenacious, tenacity, determined, determination
- hard-working, painstaking, detail, detailed, diligent
- courage, courageous, bravery, brave
- passion
- commitment (without modifier), responsibility, responsible
- intense, intensity
Journalists have mental acumen
- intelligent, smart,
- curious, inquisitive,
- knowledgeable, educated
Journalists can communicate complicated topics clearly
- clear, clarity
- communicate, write, story
- creative, innovative
Journalists provide work that is timely and proximate to happenings on the ground
- immediate, immediacy, speed, deadlines, 24/7,
- at/on the scene
- source, witness, connection
- flexibility, adaptability
- perspective
Threats and Breaches
We asked, “If you feel comfortable, please share an example of a cyber threat or breach (e.g., stealing personal information, accessing private messages, etc.) you experienced in the past year.”
This question received far fewer responses than the other four questions, and a number of the responses were comments that the respondent had either not experienced a breach or did not feel comfortable providing an example. After those non-responses were removed from the set, researchers read all responses and coded them in a bottom-up fashion. Specifically, responses were coded for the presence or absence of several forms of cyber threats and breaches, which were not treated as mutually exclusive.
- Unauthorized access. This category refers to others gaining or attempting to gain unauthorized remote access to devices or accounts (hacking).
- Accessing private messages or personal data theft. This sub-category includes only responses that specified that the hacks successfully accessed private messages or data and, in some cases, published them. (Some personal data theft occurred through breaches of consumer databases.)
- Phishing. This sub-category refers to legitimate-appearing emails or links used to collect data such as login credentials or personal information that can then be used for account access, identity theft, etc.
- Insults, threats, and harassment. This category includes a range of harassment, insults and threats in public forums (e.g., comments) and private ones (e.g., email, direct messages, etc.).
- Social media cloning. This category includes online impersonation and setting up false profiles.
- Doxing. This category refers to the release of personal information that can enable threats and harassment from third parties, including contact information and physical addresses. (In some cases, doxing was linked to unauthorized access as well.)
- Distributed denial of service (DDoS). DDoS attacks are attempts to overwhelm a website or server with fake traffic to render it inoperable.
- Other. This category includes a number of rarer incidents, such as theft of physical devices.
We also tried to identify who was targeted by attacks:
- Journalists: This category refers to cyber attacks that respondents indicated were a targeted attack against them personally; these responses may specify that attacks took place because they are journalists or it may be implied.
- News organizations: This category refers to cyber attacks that impacted a news organization’s journalistic process in some respect. These types of attacks were likely not targeted at an individual journalist specifically, but were meant to limit an organization’s capacity to produce or distribute journalism.
- Journalism incidentally: This category refers to large-scale data leaks and other cyber threats that are not necessarily an attack against journalism. The average person with a digital presence is routinely susceptible to cyber threats and some of our respondents specifically indicated that breaches were of this type.
- Not clear: Responses were coded into this category if they did not clearly fall into one of the other categories or if the intention of their attacker was unclear.
We also tried to indicate whether an attack led to a breach or not; these categories were only relevant to incidents classified under unauthorized access:
- Breaches were defined as an external actor achieving some sort of access to a person or entity’s private online information or successfully taking down a site.
- Attempts were defined as attacks that were thwarted by cybersecurity practices or tools.
- Some responses were coded as not applicable. Specifically, online threats and harassment cannot necessarily be stopped by good cybersecurity practices and they have a variable impact on the victim.
Acknowledgements
The Center for News, Technology & Innovation (CNTI) partnered with several journalism organizations who, in turn, shared CNTI’s survey with their membership. Responses were collected between October 14, 2024 and November 24, 2024 using Qualtrics’s online survey platform.
The partner organizations included:
- Centre for Journalism Innovation & Development (CJID)
- Frontline Freelance Mexico
- FT Strategies
- Global Forum for Media Development (GFMD)
- Global Investigative Journalism Network (GIJN)
- Internews
- Online News Association (ONA)
- Organización Editorial Mexicana (OEM)
- Society of Freelance Journalists
CNTI greatly appreciates the time and effort each partner gave to this project and for facilitating the survey recruitment information to their membership.
Participation was also solicited through various news and journalism-related email groups, newsletters and Slack channels — some of which included:
- Institute for Nonprofit News (INN)
- News Nerdery
- Newspack Community
- Shorenstein Center on Media, Politics and Public Policy
CNTI thanks participants for their time in taking part in a comprehensive survey about journalism.
As with all CNTI research, this report was prepared by the research and professional staff of CNTI. This report was authored by CNTI’s Research Team (Amy Mitchell, Celeste LeCompte, Emily Wright, Jay Barchas-Lichtenstein, Nicholas Beed and Samuel Jens). The work could not have been completed without our colleagues Chelsey Barnes and Connie Moon Sehat; our copy editor Greta Alquist; our graphic and web designers, Jonathon Berlin and Kurt Cunningham, as well as the team at MG Strategy + Design; and our communications team at Black Rock Group.
CNTI does not lobby for or propose specific legislation and instead is dedicated to supporting policy creation through further research and collaborative, multi-stakeholder discussions.CNTI is generously supported by Craig Newmark Philanthropies, John D. and Catherine T. MacArthur Foundation, John S. and James L. Knight Foundation, The Lenfest Institute for Journalism and Google.
Topline survey data
Below is a link to our topline data.
- Given the lack of consensus about what “Artificial Intelligence” encompasses, we use the term broadly to refer to “sciences, theories and techniques whose purpose is to reproduce by a machine the cognitive abilities of a human being.” While there is no agreed-upon technical definition, it’s helpful to consider examples like Large Language Models (LLM), which are “trained” on data to recognize statistical patterns and use those patterns to generate plausible text. These kinds of models typically have too many parameters to be fully transparent or explainable, even for their creators. ↩︎
- Full demographic tables are available in the topline. ↩︎
- There is known variability in survey responses across countries regarding social desirability and acquiescence. ↩︎
- We use the United Nations’ definitions with two exceptions: we consider Mexico and Turkey to belong to the Global South. ↩︎
- Regime type classifications come from V-Dem (p. 292). A breakdown explaining the four regime types is available here. We use V-Dem’s data for 2023, which was published in March 2024, and we collapsed the two types of autocracies into a single category both because of relatively small numbers and because constraints on freedom of expression (e.g., government censorship, media bias) are a defining trait of both types. ↩︎
- See this document. ↩︎
What It Means to Do Journalism in the Age of AI
Share
Continue Reading
-
If, When and How to Communicate Journalistic Uses of AI to the Public
Conclusions of a Day-Long Discussion among Global Cross-Industry Experts
-
Focus Group Insights #3: In a Digital World, Getting the News Requires More Work, Not Less
Defining News Initiative
-
Enabling a Sustainable News Environment: A Framework for Media Finance Legislation
An analysis of 23 policies affecting over 30 countries