在这个炫酷的科技时代，我们享有快捷便利的电子招车无缝出行服务（Mobility As a Service）、卫星定位周边功能即时显示消费好去处，以及源源不绝的临近商品价格促销。但是伴随这些便利而来的大规模数码监控和系统性算法偏见（algorithm's systemic bias），却让我们逐渐失去了实现社会流动（social mobility 注1）的能力。
为了在由科技巨头主宰的时代中生存，我们需要以一种可以取悦主流算法的方式留言、生活、工作和消费。于是，你开始察觉每个人似乎都展现同样的信念与个性、大同小异的品味，相去不远的消费方式，以及相若的价值判断，以免被算法惩罚或压制。
在这个机器算法精心安排的世界中，我们还有自由意志吗？你可曾想到过，我们过去有多少琐碎乃至关键的决策，是由旨在优化企业收益的算法所决定或推翻的？
前些时候，我对面子书网友发牢骚，人类的习性将日趋被我们过往在网络所留下，以及在未来持续通过线上活动积累的数码踪迹（digital trace）所定义、强化（reinforce）和框限，最终丧失真正意义上的自由意志（free will)。
我的一些假设（hypothesis）如下：
每个社交媒体用户，均可被一组描述个人特征的参数（parameters）所定义，例如剑桥分析公司（Cambridge Analytica）丑闻中描绘的情景。
为了有效的驱动消费，网络社交媒体有可能暗地里教导广告客户，如何通过增加或抑制能够影响我们，或我们最信任的家人与朋友之数码事件（digital events）的剂量，如传递信息的频率，和内容口味的轻重等，来达到高效的行销，或是削弱对手公信的目的。
如果以上举例的数码事件未能影响我们的信念和行为；我们身旁几同“数码奴隶”（digital slave）的亲朋戚友，可能遭到算法所产生的数码事件助推（nudge），发挥动摇我们的信念和改变我们行为的“助攻”角色。
据报道，即使一个人从未注册过面子书帐户，在技术上，面子书仍有可能会通过一个影子账户（shadow profile）来存储和追踪此非用户的数码踪迹。理论上，只要非用户的家人或朋友，甚至是陌生人上传可以辨识此人的图像、视频和帖文到社交网络上，平台可以通过机器学习算法持续拼凑和推敲这个人的日常点滴。
诚然，客制化算法的影响并非每次100％收效，也不必如此，就像行销和广告业者只图高于某个百分比的转换率（conversion rate）一样。然而，可怕的是，数码影响永远不会停止，也难以辨识边界。它会通过我们没有意识到的细微方式和隐晦渠道来形塑我们的行为和世界观。
我担心的是，由于受到企业盈利驱动数码事件潜移默化的影响，每个人都开始趋同，变得一致。
当然，在算法驱动的年代，我们生活的世界还是会有多样性，但这会是一个逻辑悖论：单元化的多样性，而不是真正的多元化。多样性的橱窗展示终究会汇集（converge）到主流品牌或超级平台（Super App）的麾下。最终的发展结果或会是，样品多元，但品牌、平台和载体极其单元。而每一个行销季节的样品主题，很有可能被超级平台的演算法逐渐单元化的设定和垄断。
另外，近日《中文哈佛商业评论》在《当算法加剧了不平等，普通人还有上升空间吗？》一文中指出，起步卑微却能白手起家取得成功的事例，在日益由人工智能驱动的社会里变得渺茫。事缘人工智能的复杂算法由少数企业精英所支配；而普通员工却得面对算法管理、指派任务、评估表现的被动命运。最终，被算法监控、几近被透支应用劳动力的员工所将面临的是晋升无望，工作两极分化的窘境。
该文作者迈克·沃尔什（Mike Walsh）在《算法领导者》(The Algorithmic Leader）一书中，探讨了令人担忧的数码阶级鸿沟，亦即我们的劳动市场被划分为三个等级：为算法工作的普通百姓、拥有设计和训练算法系统技能的专业特权阶层，还有坐拥管理算法平台的少数超级富有贵族。
该文也引用“代码上限”一词来描绘阻止普通员工在人工智能驱动的企业里晋升的现象。这现象指的是，在日常的工作环境中，初级员工和自由工作者极少与其他人类同事互动。反之，他们都是由算法管理的，也就被动的被算法精算式的利用和操控，难以突围或吸引人类高层的青睐。
今年9月，最初发表于《人物》杂志，在中国网民间广传的文章《外卖骑手，困在系统里》就刻画了系统演算法规定骑手必须完成送餐的时间、指定送餐路线，以及逾时送达食物所将面临的排名积分下跌和收入减少的悲歌，揭露了外卖骑手每日长时间受制于系统算法，牺牲可支配时间的冰冷现实。《新浪科技》分析，平台算法构成了优化资源配置效率、扩大市场份额的生命线，因此通过规则让骑手缩短送餐时间以增加接单量，是导致骑手被困在系统算法里的根本原因。
以上种种发人深思的议题，足以触发哲学性叩问：我们一直以为可以凭自己的想法做出选择，但其实在公在私，我们的选择有可能是已经被算法计算好的（并牢牢地限制住了）；在算法的时代，我们还有自由意志吗？
注1: 社会流动（social mobility）或阶级移动是社会经济学的一个概念，指的是自下而上的社会阶级升迁，例如由贫转富、由劳动阶层晋升上流社会等。
文章发表于2020年1月14日的《东方日报》
The saddest truth about this new technology era is that we have mobility as a service (MaaS) but we gradually lose out the ability to achieve social mobility due to mass surveillance and systemic algorithmic bias imposed by the technology giants and the public and private institutions who subscribe to the tech giants' one size fits all technology solution and transformation remedy.
To survive in this era that is completely dominated by the tech giants, we need to believe and behave in a way that could please the mainstream algorithms. The most scary unintended consequence is that you start to find that everyone seems to share the same belief, same taste, same spending pattern and same value judgement to be not penalized or suppressed by the algorithms, be it in the social, cultural, economic or political discourse.
Another scary thing is that the highly convenient be-all know-all SuperApp that we are using in our everyday lives is very likely to be equipped with Machine Learning capability. The users' behaviors will be modeled via Neural Net with parameters that define the human users' attributes and tendencies. The corporate will treat human users as a Neural Network whose behaviors can be "fine-tuned" and influenced towards certain belief or pattern by increasing or decreasing the dose of Nudging effect (in the form of parameters) via machine algorithms orchestrated digital events.
Do we still have free will in this machine algorithms orchestrated world? Have you ever thought that how many of our past crucial and trivial decisions were shaped or nudged by the algorithms that aim to maximize the corporate profit?
Some of my hypothesis are as below:-
Every social media user is defined by a comprehensive set of parameters like the scenario portrayed in the Cambridge Analytica scandal
The social media companies actually could have secretly taught the advertisers to increase the dose or suppress the dose of digital events that can influence us or our most trusted family and friends.
If the digital events failed to influence our belief and behaviors; our close friends and family members who have been the "digital slaves" might play the vital role to swing out belief.
Indeed there is no escape route. It was reported that even if one never ever registers a Facebook account, Facebook still stores and tracks the digital traces of this person via a shadow profile that is created and continued to be enhanced and accumulated when this non-user's family and friends uploaded the pictures and videos that contain him/her.
The micro-targeted algorithm's influence is not 100% efficient each time and it needn't to be. It's like conversion rate.
For example, for the originally neutral or anti-Chinese Communist Party folks who have been exposed to digital political brainwashing events, if after some months, some of them start to appreciate the view that China's prosperity and stability supersede human rights, freedom, equality and justice, it's a big win for Chinese Communist Party's propaganda arm already.
The horror thing is that digital influence never ever stops, and it creeps in via subtle and nuanced channels that we are not even realized.
The operating procedure of digital influence is similar to that of the marketing strategy of consumer products: there are hard selling, soft selling, online channel and offline channel.
And the effect of this digital influence is deadly pervasive, sometimes it could go beyond the anticipation of the algorithm designers and programmers who implement the logic of the algorithms.
My worry is that as a result of this creepy corporate profit driven digital influence and brainwashing, everyone starts to converge and become the same.
Of course, the main drivers of all these are algorithms driven by consumerism.
Needless to say, there will still be window dressing diversity, that are converged into the hands of very few giant platforms.
参考资料/Reference
[1] Mashable: Facebook Shadow Profiles: What You Need to Know
https://mashable.com/2013/06/26/facebook-shadow-profiles/
[2] The Conversation: Shadow profiles - Facebook knows about you, even if you're not on Facebook
https://theconversation.com/amp/shadow-profiles-facebook-knows-about-you-even-if-youre-not-on-facebook-94804
[3] 中文版哈佛商业评论: 当算法加剧了不平等，普通人还有上升空间吗https://mp.weixin.qq.com/s/YK8Oj5f7vYCbrPlnBph69g
[4] 人物: 外卖骑手，困在系统里
https://zhuanlan.zhihu.com/p/225120404
[5] 新浪科技: 被困在系统里的，还有点外卖的人
https://finance.sina.com.cn/tech/2020-10-25/doc-iiznezxr7921043.shtml
[6] 井岛俊一： 处处都是算法的时代，人类还有自由意志吗？https://www.36kr.com/p/1723471298561
[7] 陈根： 算法时代下，自由意志的让渡和进路
https://www.sohu.com/a/420468946_250147
Posts
0 Comments