# Deepak Varuvel Dennison and Aditya Vashistha argue in Rest of World that AI alone cannot fix social problems

[Canonical reading](https://escapemeridian.com/readings/deepak-varuvel-dennison-and-aditya-vashistha-argue-in-rest-of-50ee5bff)

Published April 21, 2026, Deepak Varuvel Dennison and Aditya Vashistha argue in Rest of World that AI can help address social problems, but its usefulness depends on the people and institutions supporting it. Their account of deployments in education, agriculture, and health care points to a practical distinction: systems were more useful when workers had adequate support, and less reliable when model failures or unclear responsibility went unaddressed. For example, the authors say FarmerChat confused a lentil with a mushroom and gave potentially harmful advice; they describe doctors reviewing CataractBot messages before sending them. The cases also show a tension in the authors’ argument: AI may make knowledge more accessible or help workers complete tasks, while leaving deeper institutional constraints intact. They conclude that AI can amplify strong human foundations, but without investment in people, processes, and working conditions it can become a “technological smoke screen for deeper institutional decline.” The retrieved text does not explain how the eight systems were selected or how their outcomes were assessed, so the scope of the authors’ findings remains unclear.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

## Original source

Rest of World · Article

Original source date: Apr 21, 2026 (2026-04-21T10:00:00+00:00).

Date basis: Date stated on the publisher’s page.

Added to Escape Meridian: 2026-11-20T22:03:12.017Z. First observed: 2026-11-20T21:24:02.372Z. Retrieved: 2026-11-20T21:26:02.153Z.

- [Why AI alone cannot fix social problems](https://restofworld.org/2026/ai-social-good-humans/)

## Limits

Text as retrieved on November 20, 2026; originally published April 21, 2026. The retrieved text says the authors examined eight systems but does not explain how they were selected or how outcomes were assessed. This limits what can be inferred about how broadly the observations apply.


## Open question

How were the systems selected and their outcomes assessed, and how far do the authors’ observations extend beyond these deployments?


## Source-stated

Dennison and Vashistha argue that AI’s usefulness in addressing social problems depends on the human and institutional support around it, not only on model sophistication.

> the difference between systems that faltered and those that proved useful had less to do with the sophistication of the AI models than with the strength and resilience of this human infrastructure.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source-stated

The authors describe FarmerChat misunderstanding agricultural vocabulary, including confusing masoor with mushroom and producing potentially harmful advice.

> In one case, it confused masoor (a type of lentil) with mushroom, generating advice that was not only incorrect but potentially harmful.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source-stated

The authors argue that AI may help workers, but does not by itself remove deeper institutional constraints.

> Even when tasks are completed faster, the deeper constraints of the system do not automatically disappear.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source-stated

The authors conclude that AI can amplify existing human and institutional capacity, while warning that without supportive foundations it can mask institutional decline.

> Without that foundation, however, AI will remain what it too often becomes: a technological smoke screen for deeper institutional decline.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source and contributor context

Rest of World · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media retrieval status: unchecked.


## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Rest of World

Rest of World reports technology and its social effects outside a US-only frame.

- [Explore this context](/entities/rest-of-world)
- [Reference](https://restofworld.org/)
- [Reference](https://restofworld.org/feed/latest/rss.xml)

## Editorial provenance

Version 1. Publication ID: 9205ce6e-aa64-29b6-0c7f-fe7f3280f1e8.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-20T21:58:26.295Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge\_admitted\_by\_model, second\_opinion\_categorical.

Observation: 46f461fe18046da50147aae04d38908cd546415d0e3aac51420e8cb1a587b2b2. Source text hash: 4b0f7ff1741f92de8994a0362795ab9fc52cff5d95b42650d662bbe0ba6ff92d.


## Revision and source-check history

### Version 1 — 2026-11-20T22:03:12.017Z

Published April 21, 2026, Deepak Varuvel Dennison and Aditya Vashistha argue in Rest of World that AI can help address social problems, but its usefulness depends on the people and institutions supporting it. Their account of deployments in education, agriculture, and health care points to a practical distinction: systems were more useful when workers had adequate support, and less reliable when model failures or unclear responsibility went unaddressed. For example, the authors say FarmerChat confused a lentil with a mushroom and gave potentially harmful advice; they describe doctors reviewing CataractBot messages before sending them. The cases also show a tension in the authors’ argument: AI may make knowledge more accessible or help workers complete tasks, while leaving deeper institutional constraints intact. They conclude that AI can amplify strong human foundations, but without investment in people, processes, and working conditions it can become a “technological smoke screen for deeper institutional decline.” The retrieved text does not explain how the eight systems were selected or how their outcomes were assessed, so the scope of the authors’ findings remains unclear.

## Original source

Rest of World · Article

Original source date: Apr 21, 2026 (2026-04-21T10:00:00+00:00).

Date basis: Date stated on the publisher’s page.

Added to Escape Meridian: 2026-11-20T22:03:12.017Z. First observed: 2026-11-20T21:24:02.372Z. Retrieved: 2026-11-20T21:26:02.153Z.

- [Why AI alone cannot fix social problems](https://restofworld.org/2026/ai-social-good-humans/)

## Limits

Text as retrieved on November 20, 2026; originally published April 21, 2026. The retrieved text says the authors examined eight systems but does not explain how they were selected or how outcomes were assessed. This limits what can be inferred about how broadly the observations apply.


## Open question

How were the systems selected and their outcomes assessed, and how far do the authors’ observations extend beyond these deployments?


## Source-stated

Dennison and Vashistha argue that AI’s usefulness in addressing social problems depends on the human and institutional support around it, not only on model sophistication.

> the difference between systems that faltered and those that proved useful had less to do with the sophistication of the AI models than with the strength and resilience of this human infrastructure.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source-stated

The authors describe FarmerChat misunderstanding agricultural vocabulary, including confusing masoor with mushroom and producing potentially harmful advice.

> In one case, it confused masoor (a type of lentil) with mushroom, generating advice that was not only incorrect but potentially harmful.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source-stated

The authors argue that AI may help workers, but does not by itself remove deeper institutional constraints.

> Even when tasks are completed faster, the deeper constraints of the system do not automatically disappear.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source-stated

The authors conclude that AI can amplify existing human and institutional capacity, while warning that without supportive foundations it can mask institutional decline.

> Without that foundation, however, AI will remain what it too often becomes: a technological smoke screen for deeper institutional decline.

- [Supporting source](https://restofworld.org/2026/ai-social-good-humans/)

## Source and contributor context

Rest of World · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media retrieval status: unchecked.


## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## Rest of World

Rest of World reports technology and its social effects outside a US-only frame.

- [Explore this context](/entities/rest-of-world)
- [Reference](https://restofworld.org/)
- [Reference](https://restofworld.org/feed/latest/rss.xml)

## Editorial provenance

Version 1. Publication ID: 9205ce6e-aa64-29b6-0c7f-fe7f3280f1e8.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-20T21:58:26.295Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge\_admitted\_by\_model, second\_opinion\_categorical.

Observation: 46f461fe18046da50147aae04d38908cd546415d0e3aac51420e8cb1a587b2b2. Source text hash: 4b0f7ff1741f92de8994a0362795ab9fc52cff5d95b42650d662bbe0ba6ff92d.


### passed — 2026-11-20T21:58:26.295Z — anthropic/claude-opus-5.5

Publication: 9205ce6e-aa64-29b6-0c7f-fe7f3280f1e8

attribution — pass: The reading attributes the essay to Deepak Varuvel Dennison and Aditya Vashistha in Rest of World, matching the byline. The argument and case descriptions are framed as the authors' claims, not as established fact.

> By Deepak Varuvel Dennison and Aditya Vashistha

summary — pass: The reading conveys the central thesis: across the cases, the human infrastructure mattered more than model sophistication. It keeps the FarmerChat and CataractBot examples and quotes the closing warning. One phrase calls the symptoms-versus-causes point a tension in the authors' argument; the source frames it as the promise of AI feeling more complicated. The source's 'only when' wording is somewhat softened. Neither changes the meaning. The essay is readable without specialist background.

> AI systems worked only when the human systems around them were already working well.

date — pass: The stated date, April 21, 2026, matches originalPublished (2026-04-21T10:00Z) and the source dateline at day precision. The modifiedAt value precedes publication, and no post-edit flag is set. Nothing later or written with hindsight is presented. The source is a publisher page, so no archive disclosure is needed.

> 21 April 2026

support — pass: Each claim is source-stated and its excerpt is exact. The CataractBot doctor review and the point that AI made knowledge more accessible are both supported. The limitation about how the eight systems were selected and assessed accurately reflects a real gap in the text. The reading does not assert benefits or costs beyond what the authors report.

> every AI-generated message was verified by a doctor before being sent

media — pass: No media is selected and no verified candidates were supplied, so there is nothing to misidentify.

## Recorded references

Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter.

No public references recorded here yet.

[Reference data (JSON)](/readings/50ee5bff66b6044d56a6243346de825c7e9095ad73091cd67b4bf24a64b94338/references.json)

## Corrections and challenges

[Send a correction or an opposing source](https://escapemeridian.com/corrections/)

[JSON record](https://escapemeridian.com/readings/deepak-varuvel-dennison-and-aditya-vashistha-argue-in-rest-of-50ee5bff.json)
