地铁里的 AI 伪人广告，正在集体霸凌人类
体育西路站，人称「地狱西」，是广州地铁拥挤程度的显化。
每日有数十万人聊着天、握着手机经过站厅通道，被广告灯箱照出一道道黑色剪影。
但现在，除了老板、客户和 deadline，起早贪黑、睡眼惺忪的路人们，还要面临一种新的恐惧——
那就是在地铁站里无处不在、时刻勾起恐怖谷效应的「AI 伪人广告」。
AI 味太重，广告看了很烧心
AI 的普及，向来是悄无声息的；或者说，你知道它们会火，但没想到火得那么快，好比一夜之间发现身边所有人都在看红果短剧。
地铁站的 AI 广告也是如此。
AI 广告其实不是什么新鲜事。早在两年前，深圳地铁亲自下场制作的「你的样子」系列广告，还有飞猪、京东们那些略带诡异的素材就已上过热搜。
但有 AI 广告是一回事，只剩下 AI 广告是另一回事。如今站在地铁站环顾四周，几乎全部灯箱广告里都标注着「画面由 AI 生成」，带来的是一种完全不同的精神冲击：
这一天来得比想象中快太多了。
当然，如果不是广告的「AI 味」太重，我们本不会注意到放在海报角落的 AI 生成标注。
就这一点来说，GPT Images 等常用模型难辞其咎。
每次模型更新，总有读者在爱范儿和 APPSO 的评论区提问生成图片的「脏感」有没有得到解决，最后却总是失望而归——至少在目前，未经后期处理的 AI 生成产物，还是很容易露出破绽。
但更让人难绷的是，品牌方们正在争相把这些「一眼假」的直出素材原封不动地搬上海报，一场「伪人」盛宴就此开席。
最近在广州地铁霸屏的蛋白粉广告，就可以说是「伪人味」浓到齁鼻：
一张无比典型的 AI「平均脸」、毫无纹理细节的皮肤和油腻的画面质感，足以让每个刷过抖音红果、用过 AI 生图的人说出那句「这妹妹我在哪里见过」。
甚至当我前往长沙旅游时，也没能摆脱这则广告的步步紧逼——
只是随着主打产品从蓝罐换成白金罐，人物身上的骑行服也从蓝黑换成了灰白配色，顺带让人重温了一把 11 年前火爆的「裙子是蓝黑还是白金」大争论。
代表着高端休闲需求的邮轮旅游同样未能免俗。
地铁站投放的邮轮广告视频里，三五好友手执酒杯，在甲板上相聚畅聊，一切看着都很正常……直到他们露出几乎是复制粘贴出来的大白牙。
正筹备着月底万圣节活动的长隆欢乐世界，则策划了一整套主题异空间宣传广告。
但无论是「油光满面」的人物，还是字体穿帮的异空间大门，都没有透露出半分乐园真实环境；看上去不像主题乐园，而更接近藏在北京路破旧大厦里的劣质密室逃脱海报。
从站外商场的易拉宝，到车厢内部布置，类似的例子还可以无限列举下去，毕竟现在不用 AI 生成的「古法广告」才是小众。
但看着这些广告，你很可能会跟我有同样的疑问：AI 广告，到底告诉了我们什么？
从字面意义上看，「广告」的意义应该是将重要信息广而告之，但很显然，在 AI 广告上我们并不能得到什么有效信息——
AI 人物可以代言保健用品，但他们天生就是完美的，生成的健康形象对普通消费者很难说有多少意义；车厢海报里的 AI 多巴胺少女手握一杯奶茶，但每个人都心知肚明她不可能喝过奶茶；而虚构的邮轮和乐园场景，也没有为真正可能前往现场的我们提供增量知识。
如果说传统广告希望传递的是产品、服务、品牌或观念，那么在 AI 广告身上，形式已经偏离了功能，信息几乎全部丢失，最后传递给我们的，只剩下一种担忧跟不上 AI 大潮的焦虑。
当然，我们也不应该过度神化传统广告，传统广告里敷衍了事的作品也不在少数。
最爱在电视节目中间插入的经典汽车广告，往往只是一辆车驶过荒野、桥梁和冰原，配上激昂的音乐，最后在所有观众都没有记住车型名字时就草草结束。
但不同的是，由 AI 批量生产的面孔、光线和场景，刚好构成了最易被肉眼识别的工业化痕迹。违和感与扑面而来的恐怖谷效应，几乎是生理本能地引起排斥与不适。
从这个意义上说，AI 广告不自觉地成为了一种劣质广告的「显影剂」。
AI 广告最糟糕的一面，也是它最大的贡献。挥之不去的「AI 味」，像是学术上的荧光标记，让我们意识到：原来公式化、流水线生产的无聊广告，早已遍布每个角落。
换言之，我们厌恶的也许从来就不只是 AI，而是借助新技术变得更加肆无忌惮的敷衍态度。
创意与诚意，才是好内容的秘方
讽刺的是，在被 AI 全面霸占之前，地铁站里其实不缺好广告。
去年 3 月，童装品牌 balabala 在多个城市的地铁站投放了《中国小朋友》特别摄影展。相机镜头下的每个孩子，都有自己独特的幸福与烦恼：
有人落选班级大队长，强忍着泪的眼睛里带着几分委屈与不忿；有人躲在院子里，一边窥视父母动向，一边将旧衣服藏起来，想让爸妈给自己换新装；有人在爷爷的三轮上看见军人跑过，懵懂地学着电视里的模样敬礼……
向左滑动查看更多内容
这些孩子们的细微神情，是当今 AI 还无力理解也无法重现的；但恰恰是这些细节，赋予了广告情绪的力量。
而同样是在体育西路站，去年全运会期间由 Heaven&Hell 为亨氏食品投放的番茄酱广告《想赢的番茄在亨氏》，用数十个形态各异的番茄蒂 cos 成运动员形象，好玩地兼顾了番茄的「健康」与对运动精神的支持。
毫不意外，夸赞这则广告的小红书笔记迅速破 10 万赞。那段时间走入地铁站，总能看到有路人在广告前停下脚步，掏出手机拍照——
这正是创意的价值。
以上论述很容易让人以为，AI 盛行是广告「审美降级」的直接原因。但恰好相反，过去一段时间，用 AI 辅助创意的案例也不在少数：
比如，今年 6 月，年届九旬的游本昌授权 AI 数字分身出演了一支广告短片《93 岁，我拥有了人生第一个机器人》。当时的「爷叔」身体状况已经无法胜任高强度拍摄，但在 AI 营造的世界里，他看起来精神矍铄，讲述了如厕机器人让老人能够安心起夜的简单幸福。
广告最后，还是游本昌本人亲自出镜，录制了一段祝福视频；而这则广告，也成为了这位艺术家最后的作品。
虽说「画面与广告语配音均由 AI 辅助生成」的提示从第一秒就标注在画面上，但向来挑剔的 B 站观众，这一次也并没有将矛头指向 AI。因为技术没有喧宾夺主，反而让广告得以传递它真正想表达的东西——这是一则有「灵魂」的广告。
说到底，广告是服务业，服务于品牌理念和观众的内心愉悦感；但从地铁站的现状来看，很多从业者恐怕是在用制造业的视角来生产广告，误以为视觉素材的生产数量与速度，能够取代创意与诚意。
怎样拯救公共视觉空间？
既然 AI 有能力把广告做得丰富多彩，为什么最终出现在地铁站里的，却往往是最乏味的那一种？
最直观的答案，当然是成本。
传统广告制作是一条漫长的工作链，找摄影师、租棚、请模特，甚至只是为了找到一张合适的背景素材，都需要在预算、时间、版权授权与设计要求之间反复协调。
以 Adobe Stock / 图虫创意为例，商用图库素材有不同的授权等级，如果只是发在公众号倒是好办，有现成的廉价授权套餐；但如果是用在地铁灯箱或商场门头等「户外广告」领域，授权就会变得极为复杂，必须拨打专人热线商谈户外商业授权，一张素材价格也会飙升至千元以上。
而有了 AI，一段提示词就能代替过往种种复杂的沟通与摩擦。对于需要快速交付的广告公司来说，这当然是一种巨大的诱惑。
只不过，制作和创意从来不是一回事。AI 只能保证将想法变为成品的速度，但无法保证这个想法本身值得被制作出来。
更何况，商业广告的决策权一向高度分散。
设计师可能想做得更有趣，但客户经理要保证项目周期，作为甲方的品牌市场部则要考虑领导与市场的接受程度，一旦各方意见不一，能在反复修改中活到最后的，反倒可能是那个最平庸的方案。
更微妙的是，AI 现在可以给这个平庸方案提供一种看似先进的解释。一旦用上 AI，它就成了拥抱新技术的政治正确，至于广告有没有创意，反倒成了次要问题。
如此一来，就不难理解为什么 AI 广告会迅速霸占地铁站——它正是当下最大的公约数。
但我们都知道，个人理性并不总是通向集体理性。
地铁站里的广告固然是参与各方捏着鼻子能接受的方案，但就结果而言，无处不在的 AI 伪人，正在对每天路过的上百万市民造成轮番视觉暴击，也破坏了城市公共空间的视觉环境。
对于公共空间里的视觉污染，或许可以借用法经济学的一种思路：谁能以更低的成本避免损害，谁就更应该采取行动。
普通市民如果不想被 AI 广告污染眼球，仅有的办法就是不再进入地铁站，或全程直视前方不看广告，但这几乎是不切实际的要求；相比之下，从广告投放方着手，「谁污染，谁治理」显然成本更低也更可行。
事实上，很多以公共视觉空间闻名的城市，都对户外广告作出了严格规定：巴黎的《地方广告条例》将户外招牌和广告视为城市公共景观的一部分，要求商铺门面和招牌在色彩、形状与灯光上与建筑外立面相协调，营造了我们熟悉的巴黎风情；纽约则通过《分区规划条例》，保持了曼哈顿街区外立面的融贯协调。
也许有一天，中国城市也应该制定相应的户外广告管理规范，避免粗放生长的户外广告，成为「城市审丑」的一部分。
而在那之前，至少我们可以通过社交媒体发声，尝试让品牌方听见观众的声音，直至那些感动人心的精致创意广告，重新回到视野中央。
本文由 AI宝库自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
The AI Hypocrisy ad on the subway, is collectively bullying humans.
Sports West Road station, known as "Hell West", is a significant increase in the level of overcrowding in the Guangzhou subway.
Hundreds of thousands of people talk about the sky every day, holding their handles through the halls and being photographed in an advertising lampbox with a black clipping.
But now, in addition to the bosses, the clients and the deadlines, the early, dark and asleep roadmen are facing a new kind of fear --
This is the "AI Hypocrisy Advertisement" that is ubiquitous and constant in the subway station.
AI, it's too odourful. The ads are very hearty.
The spread of AI has always been quiet; or, you know, it's fire, but I didn't think it was that fast, like one night when everyone around was watching a red fruit drama.
So is the AI ad at the subway station.
AI Advertising is not really new. As long ago as two years ago, the Shenzhen subway itself made a series of ads on "You Look" and fly pigs and the slightly weird assetes.
But there's an AI ad that's one thing, only the AI ad is another. Today, standing around a subway station, almost all of the lights are marked as "photo-generated by AI" with a completely different mental shock:
This day comes way too soon than I thought.
Of course, if it wasn't for the ad 'AI' too heavy, we wouldn't have noticed the AI generation label in the corner of the poster.
In this sense, common models such as GPT Images are responsible.
Every time the model is updated, the reader asks if the "dirty feeling" that produces the picture in the comment area of Efraim and APPSO has been resolved, and is always disappointed — at least for the time being, the AI products that are not processed by post-production are easily exposed.
However, it is even more difficult to see how branders are fighting to get these "fakes" straight out of the original asset and put them on a poster, and a "sham" feast opens.
A recent protein-puff commercial on the Guangzhou subway screen can be described as a "false man's taste" with snout:
A typical "average face" of AI, a skin with no texture details, and a greasy image is enough to let anyone who has brushed Douyin red fruit and used an "AI living map" say "where has this sister been seen?"
Even when I traveled to Long Sand, I couldn't get rid of the ad.
Just as the main hit product was changed from a blue to a platinum jar, the man's cycling suit was changed from a blue black to a gray white color, and the chords were relived by an 11-year-old fiery debate about whether the dress was blue and black or platinum.
Mailer tourism, which represents a high-end need for leisure, is also not free.
In a cruise commercial from the subway station, three-and-a-half of our best friends hang out with their wine cups and chatting on decks, everything seems normal... Until they almost duplicated pasted big white teeth.
Preparations are under way for the Halloween event at the end of the month, and a series of thematic and space-based advertisements have been planned.
However, neither the "sweet" figures nor the "sweet space door" of the font-wearing gang revealed the real environment of the semi-discretion park; it did not look like the theme park, but it was closer to the poor, secret room hidden in the old building of Beijing Road to escape the poster.
From Irabo in the out-of-station mall, to the interior of the car, there's an infinity of examples, after all, of the "old-law advertising" produced by AI.
But looking at these ads, you probably have the same question as me: AI Advertising, what did you tell us?
The word "advertisement" literally means that important information should be made public, but obviously, we don't get any good information in the AI ad --
AI figures can speak for health supplies, but they are born perfect, and the health images they generate are hard to say for ordinary consumers; the Ai Dopamine girl in a car poster holds a cup of milk tea, but everyone knows she can't have milk tea; and the fictional cruises and park scenes do not provide us with the added knowledge that we can really get to the scene.
If traditional advertising wishes to convey products, services, brands or ideas, then in AI advertising, the form has been defunctional, the information is almost completely lost, and the only thing that's passed on to us is a fear that doesn't keep pace with the great tide of AI anxiety.
And, of course, we should not overdo traditional advertising, nor should there be a small number of things in traditional advertising.
Classic car ads, most often inserted in the middle of television programmes, tend to be just a car that passes through the wilderness, bridges and ice sheets, with the virulent music, and ends when all the viewers do not remember the name of the car.
But the difference is that faces, light and scenes produced in bulk by AI just constitute the most visible signs of industrialization. The effects of the valley of terror, which comes from feelings of discord and confrontation, are almost biologically responsible for exclusion and discomfort.
In this sense, AI ads became unwittingly a "visible agent" of a bad ad.
AI's worst side of advertising is its greatest contribution. The inexhaustible "AI" taste, like academic fluorescent markers, reminds us that the old formulaic, water-flowing, boring ads are already everywhere.
In other words, what we dislike may never be just AI, but the new technologies have become more reckless.
Ideas and sincerity are the secret to good content.
Ironically, before the full-scale capture by AI, there was actually a good ad at the subway.
Last March, the children's brand, Balabala, launched a special photo exhibition on "China Kids" at subway stations in various cities. Every child under shot has his own unique happiness and distress:
People who have been cast out of class leaders, who have been forced to bear the tears of their eyes, and who have been hiding in the yard, watching their parents move, hiding their old clothes, trying to change themselves, and who have seen the soldiers running on three rounds of Grandpa's show and learn how to salute on TV...
Slide left to see more content
These children's subtleities, which AI can't understand and can't re-emerge today, are precisely those details that give the power of advertising emotions.
Also at Sports West Road, last year, a tomato sauce ad by Heaven & Hell for Heinrich's food, The Winning Tomato in Henner, used dozens of different forms of tomato cos to make an athlete's image of tomato "health" and support for sports.
Unsurprisingly, the REDnote notes that complimented the ad quickly broke 100,000. When I walked into the subway, I saw people stop before the commercials, pull out their cell phones and take pictures.
That is the value of creativity.
It is easy to think from the above that the prevalence of AI is a direct cause of the "aesthetic downgrading" of the ad. But, on the contrary, there have been few examples of the use of AI-aided creative ideas in the past:
For example, in June of this year, the late 9th of the year's trip to San Francisco authorized AI to perform an ad short film, "93 years old, I have the first robot in my life." The old man was no longer capable of high-intensity filming, but in the world created by AI, he seemed mentally ill and spoke of the simple happiness of a toilet robot that allowed the old man to wake up at night.
The advertisement ended with the personal appearance of Yumoto himself and the recording of a blessing video; this was also the last work of the artist.
Although the hint that "the image and the ad word voiceover are produced with the aid of AI" was displayed from the very first second, this time the picky B standing audience did not point to AI either. Because technology does not win, it allows advertising to convey what it really wants to express — an ad with a “soul”.
In the end, advertising is a service industry that serves brand ideas and the inner pleasure of the audience; but in the current state of subway stations, many practitioners are afraid to use the manufacturing perspective to produce advertising, mistakenly believing that visual asset production volumes and speeds can replace creativity and goodwill.
How to save the public visual space?
Now that AI is capable of making a big deal of advertising, why does it end up in the subway station, which is often the most boring?
The most intuitive answer is, of course, cost.
Traditional advertising is a long chain of work, with photographers, renters, models, and even a suitable background asset, which requires repeated coordination between budget, time, copyright authorization and design requirements.
In the case of Adobe Stock/Twistle Ideas, the commercial library asset has different levels of authorization, which would have been easier if it had been just WeChat Official Account, with a ready-made, low-cost, authorized package; but if it had been used in the field of "outdoor advertising" such as a subway lightbox or a mall front, the authorization would have become extremely complex and would have to be called to a dedicated hotline to negotiate an outdoor business authorization, and a asset price would have jumped to over $1,000.
And with AI, a prompt can replace the complex communications and frictions of the past. This is certainly a great temptation for advertising companies that need to deliver quickly.
Except, production and creativity are never the same thing. AI can only guarantee the speed with which the idea is translated into a product, but it cannot guarantee that the idea itself deserves to be produced.
Moreover, decision-making on commercial advertising has always been highly decentralized.
Designers may want to do more interesting things, but the client manager needs to ensure that the project cycle, as the brand market department on the A side, takes into account the level of acceptance of leadership and the market, and that if there is a divergence of views, it will last in repeated modifications, which may be the most common option.
Even more subtle, AI can now provide this mediocre scheme with a seemingly advanced explanation. Once AI is used, it becomes the political correctness of embracing new technologies, and the question of whether advertising is creative is secondary.
In this way, it's not hard to understand why AI ads quickly take over subway stations -- it's the biggest convention in the day.
But we all know that individual reason does not always lead to collective reason.
Advertising in subway stations is a programme that is acceptable to all involved, but in terms of results, the pervasive AI hypocrites are causing a round of visual shock to millions of citizens who pass by every day, as well as damaging the visual environment of urban public spaces.
In the case of visual pollution in public space, it might be possible to borrow a line of thought from legal economics: who can avoid damage at a lower cost should act more.
The only way for ordinary citizens who do not want to be contaminated by AI advertising is to stop entering a subway station or to look straight ahead, but this is hardly a realistic requirement; by contrast, it is clear that "who pollutes, who controls" is less costly and more feasible to start with.
In fact, outdoor advertising is strictly regulated in many cities known for their public visual space: The Paris Regulations on Local Advertising consider outdoor signposts and advertisements to be part of the city ' s public landscape, requiring that the shop and signboards be coordinated with the outside of the building on colour, shape and light, creating a familiar Parisian style; and New York, through the District Planning Regulations, preserves the integration of the outside Manhattan block.
Perhaps one day, Chinese cities should adopt appropriate regulations for the management of outdoor advertising, so as to avoid growing out-of-home advertising as part of the “urban scandal”.
And until then, at least we can speak through social media, and try to get the branders to hear the audience's voice, until those moving creative advertisements get back to the center of our vision.
Auto-fetched and translated by AI Gem Box. All rights belong to the original author.