Digital Visuality Part 1

Summary

1. Digital visuality changes what “seeing” means

The lecture argues that visual culture is no longer only about what people choose to look at. Platforms, algorithms, and AI systems increasingly determine which images are created, circulated, classified, and made visible—and to whom.

• Visuality is a historically specific way of organizing and authorizing what can be seen.
• The lecture traces the concept from Thomas Carlyle’s On Heroes (1841), where visualizing history was connected to authority, through Nicholas Mirzoeff’s account of visuality and countervisuality.
• Countervisuality is the effort to claim a “right to look back” and to represent the world from positions excluded or misrepresented by dominant visual systems.
• The lecture distinguishes between:
  • Representational images, made to depict something for a human viewer.
  • Operational images, processed by machines to perform a task, such as recognizing a face or authorizing access.
• Digital visuality extends older histories of empire and authority into algorithmically mediated settings. The lecturer emphasizes that the theory must be tested across different regions, rather than treated as a story belonging only to Europe and North America.

2. Social platforms reshape everyday visual culture

Platforms do more than host images: their formats, filters, recommendation systems, and commercial incentives shape how people present themselves and how audiences interpret images.

• Instagram helped turn the attention economy into a visual economy. Its square frames, filters, selfie conventions, and influencer culture made curated visual self-presentation routine. Its aesthetics also affect physical spaces, such as cafés, museums, and installations designed to be photographed.
• Instagram beauty filters can affect users’ self-perception. One study discussed in the lecture found that all eight interviewees felt some disconnection from their filtered faces. Participants also reported heightened attention to perceived flaws after removing filters.
• TikTok is described as encouraging an “algorithmized self”: identity shaped by repeated exposure to a personalized “For You” feed, rather than mainly by a network of chosen followers. Interviewees sometimes described the algorithm as if it understood them, while also expressing unease about how much it seemed to infer.
• In China, Douyin and beauty-filter culture create related pressures. The lecture discusses “imagination tags” and the difficulty some young women feel in imagining their unfiltered appearance as attractive. Bare-faced influencer posts under a “reconcile with plain face” trend are presented as a countervisual response.
• These examples show that similar platform technologies take different forms depending on local beauty norms, commercial practices, and censorship contexts.

3. Memes can become tools of political resistance

The lecture traces the meme from Richard Dawkins’s 1976 concept of cultural imitation to internet memes understood as groups of related variations on a shared template. It emphasizes that memes have become political tools, not merely jokes.

Examples include:

• Hong Kong, 2019: Protesters used Pepe the Frog as an emblem of frustration, reinterpreting the character in a local context despite its associations with Western far-right groups.
• Thailand: Protesters used the three-finger salute from The Hunger Games against military rule. After police used water cannon and chemical irritants, demonstrators carried inflatable ducks as makeshift shields. Images of the ducks became solidarity memes.
• Sudan, 2019: A photograph of protester Alaa Salah became an iconic image of the uprising. The “Blue for Sudan” campaign also turned the color blue into a widely shared symbol of solidarity.
• More recent youth uprisings: The lecture describes humor, dancing, absurdity, and visual juxtapositions as ways to undermine official authority and circulate protest messages. Examples include Kenyan demonstrators dancing before police, Nepali protesters playing cards during demonstrations, and Moroccan detainees smiling and flashing peace signs for cameras.

The broader lesson is that low-cost, easily remixed images and gestures can travel across language barriers and serve as forms of collective resistance.

4. Regional examples show how local context shapes platform culture

The lecture compares several regions to show that digital visuality is global but not uniform.

• India and the 2025 Maha Kumbh: The event was simultaneously a mass religious gathering, a news story, and a highly circulated social-media spectacle. Influencers and viral personalities blurred spiritual authority with online fame, prompting criticism that content creation could compromise the event’s sanctity. The lecture frames this through the religious practice of seeing and being seen by the divine.
• Indonesia: Instagram is described as a platform for young Muslim women to combine religious outreach (dawah) with commercial activity (“lucrative dawah”). Images can communicate religious instruction and moral authority, not just secular lifestyle or beauty.
• West Africa: A study of the Instagram account Everyday Africa found that images intended to challenge stereotypes could still reproduce colonial-era visual patterns. Rural or pastoral scenes, basic technology, and ageing infrastructure attracted strong engagement, creating what the researchers call “hipster colonialism.”
• Nigeria: A study of Igbo women fashion designers shows a contrasting use of Instagram: designers present their work and identity on their own terms, offering an example of countervisuality.
• South Korea: Virtual influencers such as Rozy and Lucy demonstrate how synthetic personas can become commercial celebrities. The lecture notes that these figures often embody narrow beauty ideals—such as pale skin, large eyes, and small faces—making visible the standards that filters may otherwise promote less obviously.

5. Machine vision is culturally situated and can be biased

Drawing on Jill Walker Rettberg’s work, the lecture argues that machine vision is not neutral. Technologies such as surveillance cameras and facial recognition produce different consequences depending on their social, legal, and cultural settings.

The lecture outlines five aspects of machine seeing:

1. Seeing more: Enhanced vision.
2. Seeing differently: Algorithmic interpretation.
3. Seeing everything: Comprehensive monitoring.
4. Being seen: Identification and facial recognition.
5. Seeing less: Systemic bias and exclusion.

Machines can process images and act on them without conscious awareness. This challenges the assumption that “vision” necessarily involves a human or conscious viewer.

6. Generative AI reproduces and amplifies visual stereotypes

An investigation of roughly 3,000 Midjourney images found that prompts about people and places from different countries often produced highly stereotyped results. Examples included:

• Indian people overwhelmingly depicted as elderly men in turbans.
• Mexican people frequently shown wearing sombreros.
• Nigerian homes depicted as rundown.
• Indian homes represented as temples rather than ordinary houses.
• New Delhi streets repeatedly shown with pollution or trash.
• Chinese women depicted according to Western beauty standards.

The lecture’s point is that image generators do not simply reflect culture: they can reproduce and amplify dominant stereotypes at scale.

7. Deepfakes weaken photographic trust

The lecture uses the widely circulated AI-generated image of Pope Francis in a white puffer coat to illustrate how convincing fabricated images can circulate before they are verified. Such images challenge the long-standing assumption that photographs provide evidence of events.

Indian examples include:

• A deepfake placing actress Rashmika Mandanna’s face on another person’s body, which prompted public concern about deepfake abuse.
• A fake video showing Sachin Tendulkar endorsing a gambling app, illustrating the risk of fraud.
• During the 2024 general election, AI-generated videos and avatars were used in political messaging by both opposition and government-aligned actors. The lecture notes that production time for some videos reportedly fell from days to minutes, outpacing fact-checking.
• During the 2025 Maha Kumbh, fabricated images falsely showed Bollywood celebrities and a cricketer bathing at the Sangam. Fact-checkers identified visual anomalies and used AI-detection tools to assess the images as synthetic.

Together, these examples show how generative AI can disrupt trust in images across politics, celebrity culture, and religious events.

8. DigiYatra illustrates the stakes of operational images

India’s DigiYatra system uses facial recognition, linked to Aadhaar, to authenticate passengers at airport gates.

• A traveler’s face functions as an operational image: it is processed as machine-readable data to permit or deny passage, rather than primarily as a picture for a person to view.
• The lecture raises concerns about whether enrollment is genuinely voluntary, the possibility of unequal error rates, unclear data-retention practices, and limited transparency.
• DigiYatra is presented as an example of digital visuality becoming an issue of privacy and civil liberties, not simply aesthetics or convenience.

9. Art can make machine vision’s hidden systems visible

Trevor Paglen and Kate Crawford’s ImageNet Roulette allowed people to upload photographs and see how an AI model classified them. Its often offensive or absurd labels exposed the biases embedded in training datasets and classification systems. The project helped make normally hidden parts of machine vision—datasets, labels, and classifiers—visible and open to criticism.

Central conclusion

Digital visuality describes a shift from a culture in which people mainly look at images to one in which images are increasingly produced, sorted, and acted upon by machines. Social media and generative AI are major forces in that shift, but their effects vary across cultural and political contexts. The key questions are therefore not only what is visible, but who or what decides what becomes visible, on whose terms, and under which cultural and regulatory conditions.

Speakers and sources featured

Speaker

• One lecturer, not named in the subtitles.

Scholars and theorists cited

• Thomas Carlyle
• Nicholas Mirzoeff
• Jill Walker Rettberg
• Tama Leaver, Tim Highfield, and Crystal Abidin
• Richard Dawkins
• Limor Shifman
• James C. Scott
• Eva Nisa
• Trevor Paglen
• Kate Crawford
• Lee Yunhi

Lev Manovich is not cited; the Instagram book discussed in the subtitles is attributed to Tama Leaver, Tim Highfield, and Crystal Abidin.

Studies, works, and projects mentioned

• Thomas Carlyle, On Heroes (1841)
• Nicholas Mirzoeff, The Right to Look: A Counterhistory of Visuality
• Digital Visual Culture: Theory and Practice
• Leaver, Highfield, and Abidin, Instagram: Visual Social Media Cultures
• A study of Instagram augmented-reality filters in Visual Communication
• A 2023 study of the psychological effects of AR beauty filters
• “The Algorithm and Myself” (2022), as named in the subtitles
• Studies of Douyin beauty culture, “imagination tags,” and “reconcile with plain face”
• Everyday Africa / “The Insta-Gaze” study
• A study of Igbo women fashion designers on Instagram
• Research on South Korean virtual influencers Rozy and Lucy
• Rest of World’s 2023 investigation of Midjourney images
• ImageNet Roulette (2019)
• Ways of Machine Seeing, a special issue of AI & Society
• The Birth of Computer Vision, in Critical AI
• A 2025 Nature study on computer-vision research and surveillance
• Visual Culture in the Age of Artificial Intelligence
• Waging Nonviolence’s report on memes and global youth uprisings

Cases, platforms, and organizations mentioned

• Platforms and systems: Instagram, TikTok, Douyin, Midjourney, ImageNet, Aadhaar, and DigiYatra
• Protests and movements: The 2019 Hong Kong protests; Thailand’s protests; Sudan’s 2019 uprising; and youth protest movements in Kenya, Nepal, Morocco, Madagascar, Bangladesh, the Philippines, Mexico, and Peru
• Events: The 2025 Maha Kumbh
• Images and AI cases: The Pope Francis puffer-coat image; deepfakes involving Rashmika Mandanna and Sachin Tendulkar; and AI-generated election material in India
• News organizations and publications: CBC News, Forbes, Time, Fortune, NBC News, The Boston Globe, Artnet News, The New Statesman, the BBC, The Quint, BOOM, and India Forum