Following the surge of the Ghibli trend online, social media users are now embracing a new phenomenon created by ChatGPT: the Barbie Box challenge. This trend features avatars placed inside Barbie-style boxes, showcasing an array of shapes, sizes, and even professional attire.
Participants can upload their own images and provide a text prompt that specifies their desired outfit, pose, hairstyle, makeup, accessories, box design, and overall aesthetic.
Khushi Arora, an Indian expat living in the UAE, mentioned that she engaged in the Barbie Box trend to increase her follower count.
Stay informed with the latest updates.
She highlighted that in today’s social media landscape, such challenges foster a feeling of community and play a vital role in social interaction.
Danabelle, a Filipina expat in the UAE, expressed that while the previous Ghibli trend seemed like “theft of artistry,” the Barbie Box challenge felt simply “entertaining,” despite acknowledging the potential risks associated with AI.
While there are valid concerns regarding AI, she remarked that it serves as a delightful tool. “Though I understand the caution surrounding it, much like people being wary of the ocean, it’s still enjoyable to explore,” she added.
This writer and mother noted that after seeing the trend circulating, she wanted to participate and share it with friends. Nonetheless, the process was challenging; ChatGPT required multiple adjustments to her prompt to maintain privacy and prevent the misuse of real images.
Has AI started to implement its own ethical guidelines?
When Danabelle asked the AI to create a “hyper-realistic, collectible-style boxed action figure that closely resembles the uploaded photo,” ChatGPT declined her request, citing violations of its content regulations.
When questioned about this, the AI explained that while the request was imaginative and playful, generating a realistic likeness of a recognizable person, particularly in a commercial context like a toy package, breaches its content guidelines.
However, the AI mentioned that a creative interpretation, like a cartoon-style figure, could be possible as long as it doesn’t precisely mimic the individual’s likeness.
ChatGPT then modified the request to comply with its content rules: “Create a hyper-realistic, collectible-style boxed action figure inspired by the uploaded image.”
Ultimately, it still couldn’t generate the image and revised the prompt further to eliminate personal references: “Create a hyper-realistic, collectible boxed action figure of a fierce and fabulous modern poet.”
While ChatGPT might not possess a conscience, its guidelines recognize the hazards associated with hyper-realistic AI creations and prefer cartoon-style versions instead. However, the essence remains unchanged. Does it substantially matter if the terminology shifts from “modeled exactly” to “inspired,” considering a user’s photo is still integral to the process?
‘Creating a system that reflects our image’?
According to Sam Proctor, the head of technology at Asset Integrity Engineering in the UAE, humanity has always innovated and evolved, and AI represents the zenith of human creativity. Consequently, the production of images, videos, and text is unlikely to wane in popularity.
The proliferation of AI-generated trends signifies a substantial cultural transformation rather than a fleeting moment.
“As increasingly advanced tools become readily available to the general public, and image production is streamlined, we can anticipate an ongoing cycle of visual AI movements influenced by pop culture and nostalgia,” explained C. Cemper, the founder of AIPRM, an AI prompt management firm.
The Ghibli and Barbie Box trends are merely early indicators of a broader consumerization process of AI, he added.
However, AI-generated content often faces considerable backlash since many believe it undermines opportunities for human creatives, according to Tayiba Ahmed, who heads audience engagement at Street FZC, a marketing agency based in the UAE.
As awareness grows around the perceived absences of genuine effort or creativity, Tayiba suggested that AI might be more beneficial as an auxiliary tool rather than the primary means of production in commercial applications.
“For example, it’s excellent for generating references that are hard to find online or as a foundation in brainstorming sessions,” she stated.
Does ChatGPT endanger privacy?
While these trends themselves may not be inherently damaging to user privacy, awareness is essential regarding those who might attempt to exploit it. “Typically, people, not technology, are the underlying issue,” Sam mentioned.
In contrast, Christoph cautioned that such image generation might carry serious privacy implications.
Many users might be unknowingly “providing biometric data that can be stored, reused, or even sold by various platforms,” he elaborated.
“Beyond mere facial features, AI can deduce age, gender, emotional state, ethnicity, and sometimes even location based on metadata,” he noted.
“Without actively opting out or requesting data deletion, images could persist in circulation behind the scenes,” he explained.
“If we are creating a system reflecting our own image, we should not be surprised when it mirrors the darker aspects of human behavior,” Sam remarked.
For instance, a malicious organization could design a social app solely aimed at attracting users to build a model for targeting scam-vulnerable individuals, leading to more sophisticated AI-driven scam communications, he added.
“Learning” purposes
Companies hosting AI models might retain user data for “learning” purposes to enhance the technology, Sam noted.
“We engage with email, social media, cloud-based applications for work, and countless other digital tools that accumulate vast amounts of our information in the cloud. This occurs because we have a degree of trust in these systems.”
Similarly, with Generative AI, users must trust that their data will be protected. “Generally, the major players in this field prioritize data security,” Sam stated.
He believes that these trends could be advantageous in particular situations. For example, employing ChatGPT to create an illustration for a public-facing profile picture may enhance privacy since the actual image would remain concealed.
However, Christoph contended that once someone uploads an individual’s photo, “most users are unaware it may be utilized for training future models or included in datasets aimed at entirely different objectives.”