Subbu Vincent is the director of Journalism & Media Ethics at the Markkula Center for Applied Ethics. He tweets from @subbuvincent and @jmethics. He is the convener of the News Distribution EthicsNews Distribution Ethics (NDE) Roundtable.

 

In our November 17th, 2023 meeting, we briefly discussed this set of use cases on how diverse voices, perspectives, and formats in journalism may or may not enter a news feed, and the ethical concerns that arise. Since the meeting, NDE convener Subbu Vincent took input from roundtable member Connie Moon Sehat to update each case with a deeper framework of questions, and a preamble. Other members Jennifer Wilson, Logan Smith and Tina Rosenberg also provided further inputs to the cases. 

The Cases

Case: Black Women’s Voices in “UAW Strike” feed

Case: “Election Workers” stories

Case: Transgender Candidate Danica Roem’s win

Case: July 2023 Oppenheimer News Cycle

Case: Solutions Journalism

Case: Social Video News

 

---

Case: Black Women’s Voices in “UAW Strike” feed  

Applied Ethics terms: Authenticity, Fairness to stakeholder voices (Diversity), Discourse (right to participate/be heard).

Author’s note about case development methodology: This case is not based on exhaustive research. It is for illustrating normative questions around responsible or ethical news distribution. For this case, I checked: 

1. The trending news topics (or keywords) on multiple news feed products without logging in
2. Separately used incognito on the same feed/topic to compare if the same results showed. 
3. Did not use explicitly personalized feeds like the “For you” tab in Google News, for instance. 
4. Dates: The given story’s published date is within the date range of all the stories on the feed for the matched topic. I.e. there were older and newer stories on the feed when this case was documented. 

Brief: The UAW strike has been well covered across the political and policy press and to some degree local press in Michigan. One of the few stories in the cycle this week with Black worker voices (Black women workers) is an AP story that ran at the Sacramento Observer. (Nov 7) This story is rare because it carries an intersectional selection of voices at three levels of marginalization: Women, Black, and Workers.

Distributors where this story was absent: Seekr, Google News, Ground News, Yahoo News 

Distributors where this story was present: Smart News

Other aspects: It’s an AP story, Sacramento Observer is a regional Black-led/owned publisher 

Case Questions for Discussion

1. What more do we need to discover about this case?  

1. 1. Is personalization a factor? 
   2. Could this story have made it to others' feeds and not one random observer? 
   3. Did the fact that it was an AP story with Michigan sourcing, picked up by a Black media outlet in Sacramento, CA play any role? 
   4. Who else picked up this AP News piece and did their published versions get distributed on news feeds? 

2. What should happen?

1. 1. Should news feed products "know" long-tail publishers representing marginalized groups (and self-identifying as such)?
   2. Should tech identify stories carrying intersectional voices on the marginal side of the power spectrum? 
      1. How or where should such definitions and criteria be sourced?
   3. Should tech endeavor to include such stories if the "can" part can be done accurately on grounds of "not excluding voices from the distribution"? If so, to what extent?
   4. Consequences: What are the possible problems/downsides of tech being able to do these things? 
      1. Could this risk targeting audiences based on protected characteristics rather than a general audience?

3. What can happen? 

1. 1. Can tech identify that this story has Black Women’s Workers' voices? 
   2. Can tech identify that the story is published by a reputed Black Media outlet, which is a proxy to saying this outlet is more likely to source from Black people? 
   3. Can tech identify that this is a syndicated piece from AP News with quotes not from the California region, but from Michigan, even though the media outlet publishing the piece from Sacramento, California? That would mean the issue is national (or non-local only to Sacramento, CA) and the local publisher may be adding to national discourse by picking up the piece.
   4. Some news aggregation services only surface content from publishers with whom they have licensing agreements. To include otherwise suitable long-tail publishers requires services to identify, craft agreements, and address the technical limitations of some publishers. This may be difficult to achieve at a meaningful level of scale, thus impacting diversity. How do we handle this? 
   5. Consequences: What are the possible problems/downsides of tech being able to do these things? (I.e. if the answer to “can?” is a yes).  Combine this with 2d.
      1. Could this risk targeting audiences based on protected characteristics rather than a general audience?

Do you have questions to add about this case? Send them to the author at svincent@scu.edu.

See all cases.

    

---

Case: “Election Workers” stories 

Applied Ethics terms: Authenticity (Difference between regular Opinion vs Stakeholder Opinion), Fairness to stakeholder voices (Diversity), Discourse (right to participate/be heard)

Author’s note about case development methodology: This case is not based on exhaustive research. It is for illustrating normative questions around responsible or ethical news distribution. For this case, I checked: 

1. The trending news topics (or keywords) on multiple news feed products without logging in
2. Separately used incognito on the same feed/topic to compare if the same results showed.
3. Did not use explicitly personalized feeds like the “For you” tab in Google News, for instance.
4. Dates: The given story’s published date is within the date range of all the stories on the feed for the matched topic. I.e. there were older and newer stories on the feed when this case was documented. 

Brief: Election workers around the country have been receiving threats. This news cycle with the story on fentanyl-laced envelopes. There were already stories of election workers quitting in large numbers. In the middle of this, Ms. Magazine ran an article by a former woman election worker in Colorado on why she quit and the costs of the loss of experience that are going to arise in 2024. 

Key point: This is the only article that pointed out that the majority of election workers in the US are women. These threats are impacting them more. MsMagazine is a feminist magazine (self-identifying this way.) 

Type of distribution issue: Presence/Absence. 

1. The term “Election Workers/Threats” was not a news cycle “topic” on Google News, Ground News, Seekr, Smart News and Yahoo News as of Nov 9. 
2. When you search on the apps with the keywords “Election Workers” you do get a lot of relevant stories. 
3. None are listing the Ms Magazine piece by a woman election worker in Colorado about the gender angle. 
4. A local Colorado publication’s story on the general election workers’ quitting is on the feeds. 

Case Questions for Discussion

1. What more do we need to discover about this case?  

1. 1. Is personalization a factor? 
   2. Could this story have made it to others' feeds and not one random observer? 
   3. Did the fact that it is written in the first person, and an op-ed play any role? 
   4. Did any other women (or gender-angled) election worker stories from any other publications get listed? 
   5. How is Ms Magazine categorized as a publisher in the news feed products? 
      1. Is it possible women's focused publications are categorized as "soft" news and treated differently?

2. What should happen?

1. 1. Should publishers representing marginalized groups (and self-identifying as such) be identified for such coverage in news cycles? 
   2. Should algorithms exercise their responsibility to bring diversity in perspective to their feeds? 
      1. If so, should they use sourcing distinctions likely to be consistently made by such mission-oriented entities? 
      2. If so, to what extent? 
      3. Should there be a standardization of different types of mission-oriented entities for algorithms to match news outlets to? 
      4. Should there be standardization of the connections algorithms can expect between mission-oriented entities and sourcing? 
   3. Consequences: What are the possible problems/downsides of tech being able to do these things? 
      1. Could this risk being targeted to a gender-based audience rather than a general audience?

3. What can happen? 

1. 1. Can tech identify that this story has women’s voices on for elections workers threats keyword results?
   2. Can tech identify that the story is published by a reputed women’s issues outlet, with an explicit call to feminism, which is a proxy to saying this outlet is more likely to source from women and women experts?
   3. Can tech identify that this is likely an opinion/commentary piece since it is written in the first person by an author who is bylined as a former elections officer in Colorado? This person is not on the masthead so can tech identify that this is an invited or freelance article?
   4. Can there be a standardization of different types of mission-oriented entities for algorithms to match news outlets to?
   5. Can there be standardization of the connections algorithms can expect between mission-oriented entities and sourcing?
   6. Some news aggregation services only surface content from publishers with whom they have licensing agreements. To include otherwise suitable long-tail publishers requires services to identify, craft agreements, and address the technical limitations of some publishers. This may be difficult to achieve at a meaningful level of scale, thus impacting diversity. How do we handle this?
   7. Consequences: What are the possible problems/downsides of tech being able to do these things? (I.e. if the answer to “can?” is a yes).
      1. Could this risk being targeted to a gender-based audience rather than a general audience?

Do you have questions to add about this case? Send them to the author at svincent@scu.edu.

See all cases.

     

---

Case: Transgender Candidate Danica Roem’s win  

Applied Ethics terms: Authenticity, Distribution Incentives for Inclusive Journalistic Sourcing

Author’s note about case development methodology: This case is not based on exhaustive research. It is for illustrating normative questions around responsible or ethical news distribution. For this case, I checked: 

1. The trending news topics (or keywords) on multiple news feed products without logging in
2. Separately used incognito on the same feed/topic to compare if the same results showed.
3. Did not use explicitly personalized feeds like the “For you” tab in Google News, for instance.
4. Dates: The given story’s published date is within the date range of all the stories on the feed for the matched topic. I.e. there were older and newer stories on the feed when this case was documented. 

Brief: In the Nov 7 Tuesday elections stories, the fact that transgender candidates won gained coverage, expectedly. However different outlets take different approaches to contextualizing the difficult realities for trans people. The 19thNews.org site, a news outlet with a gender politics focus had this piece. 

Danica Roem becomes first transgender state senator in the South (Nov 7) 

This article has this graph: 

“More transgender adults and youth live in the South than in any other region, according to an analysis published last year by the Williams Institute at the UCLA School of Law. Yet, in terms of policy, most of the South is an incredibly hostile place to be transgender in the United States.”

The article was absent from all the aggregators I checked. When specialized issue sites cover the same story they bring useful data and context that might otherwise get lost. 

Distributors (story absent): Smart News, Google News, Ground News, Seekr

Distributors (story present):  Seekr, carried an LGBTQnation.com piece on Danica Roem

Case Questions for Discussion

1. What more do we need to discover about this case?  

1. 1. Is personalization a factor? 
   2. Could this story have made it to others' feeds and not one random observer? 
   3. Did any other publishers carry Nov 7th election coverage on Danica Roem’s victory with this angle on majority of trans adults living in the US South? Or was this really a unique 19thNews.org piece? 
   4. How is 19thNews.org categorized as a publisher in the news feed products? 

2. What should happen?

1. 1. Should publishers representing a specific politics (e.g. gender politics) and historically marginalized groups (e.g. women), and self-identifying as such be identified for coverage in news cycles? 
   2. Should algorithms exercise their responsibility to bring diversity in perspective to their feeds? 
      1. If so, should they use sourcing distinctions likely to be consistently made by such mission-oriented entities? 
      2. If so, to what extent? 
   3. Consequences: What are the possible problems/downsides of tech being able to do these things? 

3. What can happen? 

1. 1. Can tech identify that this story has a citation on transgender adults’ demographics? The citation is to a major American university report. (UCLA) 
   2. Can tech identify that the story is published by a reputed women’s issues outlet, with an explicit position on gender politics, which is a proxy to saying this outlet is more likely to source from women and women experts? 
   3. Some news aggregation services only surface content from publishers with whom they have licensing agreements. To include otherwise suitable long-tail publishers requires services to identify, craft agreements, and address the technical limitations of some publishers. This may be difficult to achieve at a meaningful level of scale, thus impacting diversity. How do we handle this? 
   4. Consequences: What are the possible problems/downsides of tech being able to do these things? (I.e. if the answer to “can?” is a yes).  

Do you have questions to add about this case? Send them to the author at svincent@scu.edu.

See all cases.

   

---

Case: July 2023 Oppenheimer News Cycle

Applied Ethics terms: Authenticity (Difference between regular Opinion vs Stakeholder Opinion), Fairness to stakeholder voices (Diversity), Discourse (right to participate/be heard)

Author’s note about case development methodology: This case is not based on exhaustive research. It is for illustrating normative questions around responsible or ethical news distribution. For this case, I checked: 

1. The trending news topic Oppenheimer on multiple news feed products without logging in
2. Separately used incognito on the same feed/topic to compare if the same results showed.
3. Did not use explicitly personalized feeds like the “For you” tab in Google News, for instance.
4. Dates: The stories carrying Indigenous American perspectives were published within the date range of all the stories on the feed for the matched topic. I.e. there were older and newer stories on the feed when this case was documented. 

Brief: Both the movie and the discourse around it had plenty of coverage, with reviews, and critiques from different angles. It resurrected discussions in Japan. There were controversies. E.g., the Bhagavad Gita-India issue. Coverage also had the usual dose of celebrity trivia. A lot of this was evident in the news feeds on Google News, Smart News, Seekr, YouTube, Yahoo News, etc. However, a key indigenous/native perspective did not break into the feeds even though articles were available. 

Indigenous/Native American perspective on Oppenheimer 

TIME.com, an opinion article by Buu Nygren, Navajo Nation president. July 21  (Buu is an indigenous leader.) 

Axios piece, report by Russell Conteras, Axios Latino, July 20 

NativeNewsOnline.net article by their editor Levi Rickert, Native American – July 24

• About the erasure of nuclear waste damage to his people 
• Cites the TIME.com piece.

Lakota Times (Indigenous news outlet), article on Navajo people by Buu Nygren, Navajo Nation president

KSL.com had a piece on 7/26, reporting on Buu Nygren’s article 

Business Insider, 7/25, 'Oppenheimer' Leaves out New Mexicans Exposed to Radiation from the Manhattan Project, Despite Local Efforts to Contact Filmmakers

LA Times 7/26, Oppenheimer’s Test Site Wasn’t Remote. It was Populated by Hispanos and Native Americans

Trigger: A Twitter chat on this indicates LA Times has responded to a trending Tweet from Alisa Valdes' critiquing New York Times for calling the Manhattan project land "desolate" when Hispanos and Native Americans were driven out. That tweet had 13M views by 7/27.

Summary of observations 

• Several Indigenous American perspectives were published during July 21-25, 2023, by both mainstream and indigenous outlets, including by indigenous authors. 
• None of these showed up in various aggregators: Smart News, Google News/search, Bing+MSN, Seekr, Ground News, and The Factual
• BUT: "Oppenheimer" feeds were showing a ton of articles on OTHER angles and controversies. (Bhagavad Gita, Japan, etc.)
• Until July 25th the Native American angle was broadly missing
• Started showing on July 26/27 around when the LA Times story happened. 

Case Questions for Discussion

1. What more do we need to discover about this case?  

1. 1. Is personalization a factor? 
   2. Could the Time, NativeNewsOnline, and Lakota Times stories have made it to others' feeds and not one random observer? 
   3. How are Native American outlets categorized as publishers in the news feed products? 
   4. What kind of reach do these organizations have on social? Are they considered national or recognizable names, or are they considered more community-focused? Does this affect their reach?

2. What should happen?

1. 1. Should news feed products, as part of responsible distribution (contribution to public discourse) make the identification of authenticity in social/group voices a principle?
      • Indigenous outlets have a greater volume of indigenous voices.
      • Mainline outlets are running indigenous voices (lesser volume), but they offer authorship occasionally as Time did to Buu Nygren.
   2. Should news feed products consider that by not including such stories and perspectives, those voices are left out? (Consequence of a.)
   3. Consequences: What are the possible problems/downsides of tech being able to do these things?
      1. Could being identified this way reduce their reach or be used for negative targeting?

3. What can happen? 

1. 1. Can tech anticipate that because the Oppenheimer film is about the Manhattan Project, which was located in New Mexico, there may be a history of stakeholders there from marginalized groups who might be weighing in with media perspectives? 
   2. Can tech identify that the Time article on Oppenheimer, released during the news cycle, was by a well-known Native American public figure (President of Navajo Nation), and hence is likely to carry an authentic perspective of the Native American people on the Manhattan Project?
   3. Can tech identify that NativeNewsOnline and Lakota Times that published perspectives on Oppenheimer (and one of them also linked to the Time article) are both Native American-owned/run news outlets? Therefore they have higher sourcing from Native American people and experts.
   4. Could such stories from different perspectives be bundled or packaged with trending news on the topic?
   5. Some news aggregation services only surface content from publishers with whom they have licensing agreements. To include otherwise suitable long-tail publishers requires services to identify, craft agreements, and address the technical limitations of some publishers. This may be difficult to achieve at a meaningful level of scale, thus impacting diversity. How do we handle this?
   6. Consequences: What are the possible problems/downsides of tech being able to do these things? (I.e. if the answer to “can?” is a yes).  

Do you have questions to add about this case? Send them to the author at svincent@scu.edu.

See all cases.

    

---

Case: Solutions Journalism 

Applied Ethics terms: Utilitarian (results-based, greatest balance of good over harm), Common Good, Discourse (right to participate/be heard), Authenticity

Author’s note about case development methodology: This case is about a generic category called solutions journalism. The Solutions Journalism Network maintains a story tracker with nearly 16000 vetted examples. (See below.) This case is for illustrating normative questions about responsible or ethical news distribution around an established and humanly identifiable category as a whole. 

Brief: The Solutions Journalism Network has a well-recognized framework to identify solutions news stories. This is a four-criteria system. To be flagged as legitimate solutions stories on the network’s Solutions Story Tracker®, stories need to 1) report on a response to a social problem 2) offer insights from lessons learned on tackling the problem 3) show evidence of effectiveness, and 4) report the limitations of the solution(s) centered in the story. The formal rubric, which is also used by the network for training, is here. 

How do I know it is solutions journalism? 

Some challenges in this space: "Solutions-like" stories get regularly reported but usually fail one or more of the four-criteria rubric. For instance, they may not cover a response to a social problem or miss out on reporting the limitations, stay in the weeds, and not offer insights, etc. 

Research supporting solutions journalism's benefits: 

• Don’t Feed Fatalism…Put Forward Solutions Instead
• Positive News magazine sees subscriber growth as editor says people don’t want journalism that leaves them ‘completely hopeless’

Examples of recent solutions journalism announcements/rollout:

• How an innovative Armenian platform battles news fatigue with solutions journalism
• How Black journalists are incorporating a solutions approach in their newsrooms

Case Questions for Discussion

1. What more do we need to discover about this case?  

1. 1. Have any product teams already tried identifying or building signals for solutions journalism? 
   2. A UT Austin study posted promising results (with caveats) on solutions journalism example getting more engagement also affecting attitudes. Would this be a starting for making a pro-business argument to surface solutions journalism articles within news cycles? 
   3. Higher-level empirical questions: 
      1. In a breaking news cycle on a mass shooting, would bringing in a contextual, high shelf-life solutions story on local gun regulations, that might have been done 3/6/12 months ago be useful for news readers? (And should such evergreen stories come with some kind of time stamp warning?)
      2. When there is extreme trauma in the news, can elevating solutions journalism on newsfeeds meet the moment? 

2. What should happen?

1. 1. Should tech try to identify legitimate solutions stories within trending news cycles? (Applying the rubric) 
   2. Should labeling of solutions stories on newsfeeds be done to make it easy for readers to select? (Labeling done by curators or human reviewers or machines, if accurate)
   3. Should evergreen solution stories come with some kind of time stamp warning?
   4. Consequences: What are the possible problems/downsides of tech being able to do these things? 

3. What can happen? 

1. 1. Can tech accurately identify legitimate solutions stories, i.e. without mistaking imposter stories? (Imposter stories will fail one or more elements of the rubric.)
   2. Can a tech-human hybrid be implemented where an editorial team reviews likely stories and flags them internally for confirmation? (Similar to a negative content moderation escalation. See Q 2b.) 
   3. Consequences: What are the possible problems/downsides of tech being able to do these things? (I.e. if the answer to “can?” is a yes). 

Do you have questions to add about this case? Send them to the author at svincent@scu.edu.

See all cases.

    

---

Case: Social Video News

Applied Ethics terms: Discourse (right to participate/be heard), Journalistic Sourcing

Author’s note about case development methodology: The examples is this case are not based on exhaustive research. They are for illustrating normative questions around responsible or ethical news distribution. 

Brief: We’ve listed several YouTube channels and TikTok accounts below. There are differences between these channels/accounts/sources and they are all doing “news”.

YouTube

WSJ’s reporter Shelby Holliday

Channel:  Shelby Holliday (English)

Comment: This is "formal" journalism - a mainstream publisher-reporter's channel. It looks like other WSJ reporters have their own channels. 

Johnny Harris, award-winning independent explainer journalist

Channel: Johnny Harris  (English)

This too is "formal" journalism, but an independent journalist's channel. He has credits and credentials in the journalism industry. He's following a genre called "explainers", and offers transparency about sourcing, credits, etc. Such work necessarily includes framing, fact selection, interpretation, and editorial judgment. Nothing unusual. This seems like a YouTube success story for independent journalism.

Ravish Kumar, Hindi

Channel: Ravish Kumar (Hindi)

This too is "formal" journalism, and is a mix of facts, factual baselining, and left-leaning commentary. There is also questioning of the current establishment. (Common journalistic value.)  It is not different from an anchor's show on Cable TV, except Ravish Sharma had to leave NDTV a year or so ago and decided to come back independently on YT. (Control of the NDTV company changed hands to new owners seen to be closely connected to the Modi administration in New Delhi.) 

Sarthak Goswami, Hindi

Channel: Sarthak Goswami (Hindi)

This is the most interesting. This creator is using current news as a baseline to mostly entertain, a-la American comedy journalism. But people may get "facts" and "news" from these episodes too, given the public sphere is very decentralized and fragmented otherwise. However, it has this disclaimer right up front and that's easy to miss because it goes so quickly. Caveats out the entire thing as satire.

TikTok

Traditional news 

“BBC News, The stories that matter to you”

Comment: This is "formal" journalism - a mainstream publisher's channel.

MSNBC: “We’re that network with the anchor who provides legal analysis via rap lyrics.”

(content created just for TT)

Comment: This is "formal" journalism - a mainstream publisher's channel.

 

Individual journalists who work for large media entities

Max Foster/CNN: “News Journalist”

Comment: This is "formal" journalism - a mainstream publisher-reporter's channel.

Sandra Gathmann/Al Jazeera: “I explain world news on a YouTube show called ‘START HERE’”

Comment: This is "formal" journalism - a mainstream publisher-reporter's channel.

 

Independent journalists 

Dylan Page: “𝗡𝗘𝗪𝗦 𝗗𝗔𝗗𝗗𝗬 #1 News Account On Tiktok”

Sophia Smith Galer: “journalist, author, languages nerd”

Gully Burrows: "stories i hope you’ll like 🎥📝 british journalist in abu dhabi"

Jules Terpak: “Exploring digital culture”

 

Citizen journalism - hyper-local content 

“Noticias en Huancayo: noticias al día de Huancayo, Valle del Mantaro y nacional” (Peru) “News in Huancayo: up-to-date news from Huancayo, Mantaro Valley and national”

Graham Smith: “Old-school hack re-inventing local journalism” 

 

Case Questions for Discussion

1. What more do we need to discover about this case?  

1. 1. Are there examples of channels/accounts/individuals who call themselves journalists, title their accounts or product or content stream as “news” or “news”-related, but are systematically airing misinformation and disinformation? 
   2. What about the influencer trend cropping up around disasters and true crime? I'm thinking of the daily updates on the Titan sub or true crime stories from people who built their following on these stories. What differentiates them from "fans" vs "journalists" and how do people know whether or not to trust them?
   3. What definition of “news” are news feed products using implicitly or explicitly? 
   4. What definition of “journalism” are news feed products using implicitly or explicitly? 

2. What should happen?

1. 1. How should algorithms treat the zone between formal and informal "journalism" with varying shades of formality, entity type, and persona? What harms must be considered?
   2. How should algorithms treat news commentary in comedic or satirical form?
   3. Who gets to use a satirical caveat, and what does it buy for the creator on the platform? Is the status the same as a journalistic news publisher but with fewer risks?
   4. Many accounts already identify as “news” or “journalists”, not necessarily both. Should any account claiming to offer news be automatically subject to some journalistic standards by policy or algorithms? 
      1. Should newsworthy claims be fact-checked in real-time by a claim-matching system?
   5. Should product teams use a definition of news offered by the standards community? 
   6. Should product teams use a definition of journalism by the standards community?    
   7. Consequences: What are the possible problems/downsides of tech being able to do these things? 
      1. Could a standard definition of news actually exclude smaller or more diverse publishers?

3. What can happen? 

1. 1. Can tech identify that newsworthy claims are being made in a video? 
      1. Can such claims be fact-checked in real-time by a claim-matching system?
   2. Can tech detect whether an independent creator is hosting the show in the video? (I.e. the person is not affiliated with a formal news organization) 
   3. Can tech detect if the creator/host has any journalistic attributes? (Professes being a journalist, had a career at a journalistic outlet before becoming independent, etc.) 
   4. Consequences: What are the possible problems/downsides of tech being able to do these things? 

Do you have questions to add about this case? Send them to the author at svincent@scu.edu.

See all cases.