Study Shows Online Images Fuel Gender Bias as Visual Consumption Skyrockets

Published

3 months ago

November 4, 2024

RTX

A new study highlights how the deluge of online images may be subtly shaping our perceptions, revealing that search engine imagery is reinforcing gender stereotypes in a way that text-based searches do not. As more than 6 hours a day is spent online on average, visual input from social media feeds, websites, and digital ads contributes to a cycle that may be increasing implicit biases in society.

The recent research analyzed image search results for various occupations on platforms like Google, Wikipedia, and IMDb. It found that images were overwhelmingly gendered, particularly in fields historically viewed as male- or female-dominated. For example, searches for “heart surgeon,” “investment banker,” or “developer” predominantly returned images of men, while terms like “housekeeper” and “nurse practitioner” were associated almost exclusively with women.

The study went beyond just measuring bias in search results. Researchers conducted an experiment where 423 U.S. participants used Google to search for occupations, with some participants receiving visual representations through Google Images while others used text-based Google News searches. Those exposed to image results displayed a marked increase in implicit gender biases, as measured by a standard association test, even days after the experiment. The findings highlight the impact of image-heavy platforms like Instagram and TikTok in normalizing biased visuals, raising concerns that the sheer volume of gender-stereotyped images might be entrenching outdated perceptions.

Vicious Cycle of AI and Bias

The problem extends to AI models, which are trained on vast repositories of online content, including stereotyped images. When users ask AI platforms like ChatGPT to visualize various professions, they often receive outputs that reflect existing biases. A request for images of “doctor” or “scientist,” for example, may yield predominantly white male figures, reinforcing societal stereotypes. Similarly, descriptors such as “successful” or “smart” also skew heavily towards images of white men, reflecting the biases embedded in the data used to train these systems.

The study’s authors warn that this cycle could worsen as AI tools continue to rely on biased online content. “The rise of images in popular internet culture may come at a critical social cost,” they write, noting that these biases not only influence AI outputs but also shape user perceptions. The more biased imagery we encounter, the more normalized these stereotypes become, perpetuating a feedback loop of implicit bias.

Seeking Solutions and Reclaiming Visual Space

Amid growing concerns, experts point to several solutions to mitigate the spread of biased visual content. Technology companies bear much responsibility, though attempts to address the issue have sometimes led to overcorrections. Google’s AI tool Gemini, for instance, has been criticized for inserting diversity where it historically wouldn’t exist, sometimes producing historically inaccurate imagery. Yet, even with the best intentions, fixing ingrained biases remains a challenge for tech firms.

One approach recommended for individuals is to curate their social media feeds to follow diverse creators and photographers from around the world. Another option is the “digital detox,” as outlined in art entrepreneur Marine Tanguy’s The Visual Detox: How to Consume Media Without Letting It Consume You, which suggests limiting screen time and reclaiming time away from devices. Tanguy advocates setting daily app timers, deleting unused apps, and spending time outdoors to reduce reliance on screens.

Perhaps most importantly, experts stress the value of self-awareness in understanding how digital imagery influences our beliefs and perceptions. Unlike previous generations, modern users encounter a constant stream of images that subtly shape their worldviews. For much of human history, art and visual media were limited, yet today’s image-saturated environment is altering how we see others and ourselves, often without conscious realization.

As visual culture continues to expand online, recognizing these subtle influences may be crucial in building a more balanced and less biased digital world.

Technology

Chinese Memes Take Over American Social Media Amid Rising Tech Tensions

Published

1 day ago

January 29, 2025

RTX

Despite growing tensions between the United States and China over national security and trade, Chinese internet culture is experiencing an unexpected surge in popularity among American social media users. From viral memes about industrial chemicals to an influx of U.S. users on Chinese platforms, the digital space has become an unlikely bridge between the two nations.

One of the most surprising trends of 2024 was the rise of Donghua Jinlong, a Hebei-based manufacturer of industrial glycine, an amino acid used in food production. Though an obscure company, its advertisements on TikTok unexpectedly went viral, transforming the brand into an internet sensation. American users, dubbing themselves “glycine girlies,” embraced the meme, creating influencer-style content about a product they had no actual use for.

Diana R., the creator behind the popular meme account CitiesbyDiana, explained the phenomenon. “We spent an entire month pretending we were buying industrial chemicals from China and making videos as if it were a luxury brand,” she said. The meme became so widespread that it was covered by major publications, including The Washington Post and The New Yorker. Donghua Jinlong even played along, sending merchandise to online creators.

TikTok Ban Sparks Migration to RedNote

The viral fascination with Chinese memes coincides with a new wave of American users flocking to RedNote (Xiaohongshu), China’s version of Instagram. This sudden migration was triggered by the brief implementation of a TikTok ban over national security concerns, before President Donald Trump granted the platform a 75-day reprieve.

In just a few days, around three million Americans joined RedNote, integrating into a platform with 300 million predominantly Chinese users. The result was an unexpected cultural exchange, as users swapped jokes, shared memes, and even helped each other with homework.

American internet users have long speculated about what lies beyond China’s “Great Firewall”, which restricts access to many Western social media platforms. But RedNote has emerged as one of the few Chinese apps accessible overseas, allowing U.S. users to engage with Chinese netizens in a way that was previously rare.

A Cultural Exchange Beyond Politics

Even as U.S. leaders push for restrictions on Chinese tech companies, internet culture appears to be bridging the geopolitical divide. One of the most popular viral phrases to emerge from RedNote is “you swan, he frog,” a Chinese idiom originally used to describe relationships with mismatched expectations. After being posted by an American influencer, it quickly spread across English-speaking social media, applied to everything from celebrities to politicians.

The rise of Chinese memes in the U.S. reflects a growing curiosity about Chinese internet culture, despite political narratives urging division. Jianqing Chen, an East Asian languages and cultures researcher at Washington University in St. Louis, believes social media plays a crucial role in fostering connections between everyday citizens. “Grand political narratives lose their power at the level of everyday interaction,” he said. “Personal engagement online allows people to see beyond tensions and recognize shared human experiences.”

Even amid discussions of trade wars and national security threats, social media continues to show that humor—and memes—transcend borders.

Technology

DeepSeek’s Rise Shakes AI Industry, Sends Shockwaves Through US Tech Sector

Published

2 days ago

January 28, 2025

RTX

A new AI app from China, DeepSeek, has sent shockwaves through the tech industry, markets, and the prevailing belief in American dominance in artificial intelligence (AI). The app’s meteoric rise has raised questions about the future of the AI sector, as well as the cost-effectiveness of AI development.

Venture capitalist Marc Andreessen referred to the launch of DeepSeek as AI’s “Sputnik moment,” drawing comparisons to the Soviet Union’s launch of the Sputnik satellite, which triggered the space race. Over the weekend, DeepSeek became the most downloaded free app on Apple’s US App Store, and by Monday, it was already having a profound effect on the markets. The app’s launch sparked a massive sell-off of major tech stocks, including a 17% plunge in the share price of AI chipmaker Nvidia, which lost nearly $600 billion in market value — the largest drop in US stock market history, according to Bloomberg.

What has stunned Silicon Valley is the claim by DeepSeek’s developers that the app was built for just $5.6 million — a fraction of the billions spent by AI giants like OpenAI, Google, and Anthropic. While OpenAI spent $5 billion in 2023 alone, DeepSeek’s developers suggest that this low cost could signal a breakthrough in AI development, potentially disrupting the established industry.

In a growing sense of disbelief, many in the tech world have been left questioning DeepSeek’s viability and the accuracy of its financial claims. Some analysts have speculated that the app’s success could indicate a shift in how AI is developed and delivered, with lower costs and fewer resources than previously thought necessary. DeepSeek’s models reportedly use open-source software and existing technology, which has led some to speculate that its use of Nvidia’s H800 chips may have been facilitated by stockpiling during the past years.

The development has also spurred concerns about US dominance in the sector, as the country has long assumed its control over high-end chips and data centers gave it an advantage in the AI race. However, China’s sudden emergence as a serious player with DeepSeek has led many to reconsider the global balance of power in AI.

Billionaire Sam Altman of OpenAI, who has largely stayed quiet on the matter, did comment late on Monday, acknowledging DeepSeek’s “impressive” performance. While Altman reaffirmed OpenAI’s position as a leader in AI, he welcomed the competition, calling it invigorating.

DeepSeek’s arrival has left many observers unsure of what the future holds for US semiconductor giants and raised broader questions about the long-term implications of AI development. With this new development, China has flexed its muscles in the AI sector, making clear that the race for dominance is far from over, and the US may no longer have a clear upper hand.

Technology

What We Can Learn From Communities That Carry Heavy Loads

Published

3 days ago

January 27, 2025

RTX

In rural Vietnam, farm workers carry loads that often exceed their body weight, walking miles daily with impressive ease. Their secret lies in the use of springy bamboo poles, which allow them to adjust their gait so that the weight oscillates with their strides. According to a study, this technique reduces the effort required to lift the load by about 18%.

While the bamboo poles play a key role, years of physically demanding labor have also made these workers exceptionally strong. For most people, carrying such heavy loads is unimaginable, but the methods and strength of these farm workers offer lessons for building strength and resilience in daily life.

Strength has fascinated humans for thousands of years. Evidence of early weightlifting competitions can be traced back to ancient Egypt, Greece, and China, with records of resistance lifting as early as 557 BCE. In modern times, weightlifting has evolved into an Olympic sport, showcasing feats like Georgian weightlifter Lasha Talakhadze’s record-breaking 267kg clean and jerk at the 2021 World Weightlifting Championships.

While such extreme displays of strength may be beyond reach for most, strength training offers tangible benefits for everyone. Regular strength training improves health and mobility, particularly in older adults, and is associated with lower mortality rates from conditions like cancer and heart disease. Research also suggests that strength training can benefit mental health, boosting overall well-being.

Developing the capacity to carry heavier loads starts with building muscle, particularly in the core and auxiliary muscles that stabilize the body. “Our bodies are adapted to carrying our own body mass, but adding extra weight requires consistent strength training to grow your load-carrying muscles,” explains Jeffrey Ackerman, a teaching professor of mechanical engineering at the Colorado School of Mines.

However, muscle size alone doesn’t guarantee strength. Training muscle activation—learning to maximize the efficiency of muscle use—is equally important. Experts caution against jumping straight into heavy lifting. Instead, they recommend focusing on proper technique and gradually increasing weights over time to build strength safely and effectively.

The lessons from Vietnam’s farm workers and other strength-building practices worldwide highlight the remarkable adaptability of the human body. With the right techniques and consistent effort, many of us can develop the strength to carry more—both physically and metaphorically.