100 Designs a Day? Only AI Can Do That!
-
The excessive demand of '100 designs a day' has sparked outrage among designers towards clients. While exploring why this demand is so unreasonable, perhaps we can also see if there are others who can take on such a 'heavy responsibility'.
Remember Wang Xiaoqin, the founder of Shenzhen Feiyue Travel, who demanded her employees produce '100 designs a day'?
Due to subsequent explosive news, this incident was quickly forgotten by netizens. A week later, she and her company seemed largely unaffected.
Completing 100 designs within normal working hours leaves an average of only 3 minutes per design—an impossible task for any human.
—But AI can do it.
Alibaba's AI design application 'Luban', renamed 'Luban' in April 2018, opened its capabilities to the public. The retail price for finished images is 10 yuan each; bulk discounts are available for large clients.
Similar competitors like ARKIE exist in the market. Meanwhile, Microsoft's Xiaoice provides a textile design platform for apparel companies, enabling on-demand creation.
AI's batch generation of creativity isn't limited to posters or long images.
Microsoft Office 365, WPS 2019, and Smartisan TNT all offer 'design inspiration' or similar features, automatically generating slides from given text and images. Tools like Weibo Cloud Clip, Tencent's AI auto-editing tool, and AlibabaWOOD provide the ability to quickly produce short videos in bulk.
Since Boss Wang's two human employees were so unproductive and even leaked their chat logs online, causing a scandal, it seems only natural to replace them with AI. AI can work 24/7 without complaint, doesn't require social security contributions, and won't start a public account to criticize you.
Would Boss Wang consider such a great deal?
After careful consideration, the editor feels things aren't that simple.
The first reason Boss Wang might avoid using AI for design is that she likely 'looks down' on AI's creative output. No matter how refined, the moment it's known to be machine-generated, it loses credibility.
Although aesthetics as a discipline has developed long-standing academic standards for judging beauty in art, humanity's collective sense of beauty remains a mystery.
In the TV series Sex and the City, a gallery visitor mistook a fire extinguisher on the wall for an artwork.
In reality, since Marcel Duchamp signed a store-bought urinal in 1917, turning it into the famous artwork Fountain, similar incidents have never stopped.
In 2015, a visitor dropped a glove on the floor at MoMA, and passersby carefully avoided it, thinking it was an exhibit. That same year, cleaners at an Italian museum discarded two artists' installations, mistaking them for trash because they 'looked' like empty bottles.
In December 2019, Italian artist Maurizio Cattelan offered the ultimate solution: he sold a banana duct-taped to a wall for $120,000, declaring that if the banana was eaten, replacing it would keep the artwork intact.
These examples reflect the chaos and self-contradiction in human aesthetic standards, showing that humans can't definitively determine artistic parameters—let alone authorship—unless the artist imposes a strong personal style.
(But AI can replicate styles too. Alibaba's 'Luban' launched the 'Reindeer Project', inviting designers to train its AI to reach the skill level of an Alibaba P6 designer by 2018's Double 11 shopping festival.)
In January 2007, Grammy-winning violinist Joshua Bell disguised himself as a street performer in a Washington subway station. Despite his virtuosity, only 7 out of 1,100 passersby stopped to listen, and he earned just $32.17. His concert tickets cost at least $100, and his performance fee averaged $1,000 per minute.
In October 2019, British YouTuber Taz and his untrained friends doodled some paintings for a gallery exhibition. The parody exhibit was taken seriously, with viewers analyzing 'clear influences from Degas'. They sold 8 pieces for 375 euros (donated to charity).
This explains why Microsoft's Xiaoice went undetected at a Central Academy of Fine Arts exhibition under the pseudonym 'Xia Yubing', and why audiences on CCTV's Smart Beyond couldn't identify which poster was designed by Alibaba's 'Luban'.
Humans, 'embarrassed and angry', overreact to AI creations, subconsciously dismissing them as 'soulless', 'just filters', or 'lacking rhythm'.
The clearest example of human vanity isn't even from art. In December 2019, Xi'an Academy of Fine Arts introduced a stir-fry robot that cooked 150 dishes a day, indistinguishable from human-made ones. A student told reporters: 'Standardized cooking lacks the chef's emotional touch, so it has no soul.'
—No soul? You still ate it, didn't you?
The second reason Boss Wang might avoid AI is that she can't assert her authority or feel the 'company's hustle' without berating subordinates.
Feiyue Travel's co-founder Jia Fei said in an interview:
'This is normal. If we're unsatisfied with 3, 5, or 10 designs, we ask for 100 to choose from. Employees can search, download, or adjust formats. If it's too much, we discuss it. No big deal.'
Clearly, the bosses agree on what's a 'reasonable workload'. But whether it's 10 or 100 designs, it's only the start of the employees' nightmare.
The boss's indecision stems from a lack of aesthetic knowledge—unable to pinpoint flaws, they can't identify the best either.
Frustrated, Boss Wang might say:
'Look at these designs—I don't even know what you're doing. This isn't a garbage dump. Do you have any basic sense of aesthetics?'
Employees, fearing reprisal, wouldn't dare ask what exactly is wrong.
Graciously, the boss offers a clue:
'Stop with the cartoonish designs—are we running a kindergarten?'
Normally, bosses would specify disliked styles upfront to avoid wasted effort—but that's too much to hope for.
Long-term employees might deduce preferences from past outbursts to avoid scoldings.
The problem is, the boss's tastes may shift with time—or mood.
If Boss Wang learns before work that her child behaved exceptionally well and scored 100 on a test, or conversely, had a fight with family members, her reactions and choices when reviewing candidate designs may vary significantly.
However, if Boss Wang admits that her selection is no different from a random computer-generated choice, it would be utterly humiliating—not just for her, but even the company president would find it unacceptable when imagining themselves in her position.
So, what's the solution? Simply reprimand subordinates.
As an employee under Boss Wang specializing in design—though "designer" is too lofty a title for this role—the key skill isn’t actually creating designs but deciphering the boss's psychological cues.
Whether designing graphics, producing videos, drafting architectural plans, writing academic papers, or crafting paid promotional content, professionals across industries interacting with clients have learned painful lessons.
For instance, when delivering an article, a client might request five draft headlines to choose from. Offering what you consider the single best headline is counterproductive. Even proposing five excellent options is unwise because clients will instinctively critique submissions to assert their authority—either pointing out flaws or suggesting changes (even if unnecessary).
Understanding this psychology, you should include your preferred headline among several options while intentionally leaving flaws in the others. When the client selects, they’ll likely choose your intended option while feeling satisfied with their "decision-making."
AI, lacking this human insight, simply provides multiple options, dumping both decision-making responsibility and accountability on the boss without offering them a chance to vent frustration. How could this please any boss?
This leads to the third problem with replacing humans with AI: accountability. When things go wrong, someone must answer—but AI isn’t human and thus can’t take responsibility.
Uber’s fatal autonomous vehicle accident, where the system misidentified a jaywalking pedestrian as an object, exemplifies this dilemma. Although a human driver was present during testing, prolonged safe operation led to complacency.
Ultimately, Uber assumed responsibility, suspending tests to investigate. But ethically, real-world deployment differs from testing. When mass-produced autonomous vehicles crash, can manufacturers bear all blame? Debates continue among AI ethics committees regarding whether owners, passengers, pedestrians, manufacturers, or governments should be liable.
If Boss Wang uses AI like Alibaba’s Luban or Microsoft’s Xiaoice for designs, the final selection rests with her. Whether she wants "multicolored black" or "smaller yet bigger," given time, AI delivers. But should a busy boss spend time on this?
Thus, AI adoption might require retaining human subordinates to serve as gatekeepers—enduring reprimands while skillfully pleasing the boss.
Accountability also involves copyright. AI-generated works’ ownership remains contentious. The U.S. Copyright Office mandates human authorship, while the European Patent Office rejects machine inventors. A 2020 Chinese ruling awarded Tencent damages for copied AI-written news, but the tool was proprietary, unlike public AI services.
With major tech firms offering public AI design tools, copyright claims remain uncertain. If competitors reuse designs without consequence, Boss Wang would surely fume.
In conclusion, hot-tempered Boss Wang is unlikely to replace loyal staff with AI. While AI provides correct answers, it undermines the boss’s authority. You might assume companies with such leadership are doomed, but Fly Over Travel thrives serving multilevel marketers—offering luxury backdrop rentals and celebrity-access services for social media clout.
Such businesses exploit human psychology and desires, requiring human employees who master flattery and avoid backtalk. In an AI-dominated future, these roles survive by catering to bosses’ egos.
Here’s hoping no human ever meets such absurd job criteria again.