Nowadays in the global context of modern technology, AI has been universally infiltrated no matter online or offline and gradually gets to know humans comprehensively in our daily lives. AI, which is artificial intelligence. It is a machine made by humans but is programmed beyond human thinking and able to do tasks more conveniently and smarter. Technically speaking, in the book Atlas of AI written by Kate Crawford. “Professor Donald Michie described AI as knowledge refining, where “a reliability and competence of codification can be produced which far surpasses the highest level that the unaided human expert has ever, perhaps even could ever, attain.” The creation of AI may exceed the knowledge of experts one day which is the developing technology human is expecting but the concerned risks are also beneath it.
According to a recent poll, “more than 72 percent of Americans are concerned about machines doing many human tasks in the future.“
Eray Eliaçık
Firstly, let’s take an example. If I use two different search engines, one is Google that I have used very frequently and the other one is Bing I rarely use. I Type in “Workout” as the same content in the search bar and see whether they all generate similar results or any differences. When I finished typing the word, the Google search bar popped up a drop-down list that included the previous records of my workout-related search and Bing didn’t. On the result page, apparently they both seem similar as the most relevant content will be shown in the first place by order, such as the definition of workout, the workout beginner guides and plans, the images of workouts, Bing even show me workout-related movies. On the other side, if you pay attention to the arrangement and presentation of search results. It is easy to find out that Google which I used a lot has a higher degree of personalization and customization than Bing. the results would present depending on my geographical location, searching records and devices and so on. As I have created an account on Google so it is allowed to locate me and provide gym locations in Shanghai. Also, if I scroll down to the bottom, there are several relevant search topics in Chinese just to ensure a more appropriate way of understanding what workout is as multilingual.
Besides, the input of keyword also varies in the algorithm of search engines. For example, our search habits, the purpose of information acquisition, the details of search terms, etc. Each of us has different expectations and preferences for search results. So in my case, If I type in “workout” on Google search, the search results will be weighted towards searching for fitness workout methods and planning, with only providing a few basic information about gyms or exercise equipment. For instance, I searched quite frequently during the pandemic about how to use a pair of dumbbells to train my entire body at home. Then the algorithm of the engine took that in mind and keep providing this kind of guide as a priority. It is generated based on thousands of related inputs and all I need to do is type in “workout”. As a result, from this little testing that indeed the more data it collects, the better results it will show mean while less effort to spend.
In fact, AIs almost can already substitute for humans to live life, more than that even considered beyond human thoughts. The reason why AI can insight into humans and know us so well is mainly because of its training data that helps algorithms to learn. However, this can unconsciously raise doubt among people over time about whether we can trust and rely on AI algorithms entirely. Are there any potential threats that we haven’t noticed yet at this stage? I argued that until now there are indeed lots of benefits of precise algorithms to humans, but the hidden problem behind it is also emerging unnoticedly and shaping our lives.
The storm spreading of Tiktok as the leading social platform app
TikTok is a short-video social software for musical creativity incubated by its mother company ByteDance in Beijing, China. The software was launched on September 20, 2016, and is a short video community platform for all ages, through which users can select songs, shoot music works to form their own works.
“Currently TikTok covers more than 200 countries and regions around the world and is ranked number one in 141 countries worldwide. As of 2023 TikTok has exceeded 4 billion total downloads in Google Play and App Store worldwide.”
Tiktok Xiaoke
So, is TikTok an AI? Although TikTok may seem like another fad or like a popular trend, “the app uses the latest technology—specifically, artificial intelligence (AI) and machine learning (ML).”(Kohen, 2022) Now telling you the truth, every step you click on Tiktok is interacting with AI progress. The moment you open the app and the algorithm generates the most relevant content to you according to your viewing habits, or every time when you hit a like or leave a comment below the video, the machine is learning and memorizing your interest trends.
How Tiktok took control of Users technically
What is algorithm? Referring to what Terry Flew defined in his book, “the computational processes through which user inputs interact with data sets in order to generate outputs.”(Flew, 2021) Over time, algorithms can be improved by studying how to react more adequately to their input from iterative engagement with users and data. But in the uncultivated wasteland of the past digital era, “Inevitably, it is difficult to grasp how they operate and how changes manifest themselves, as an algorithm of this nature is like a black box: opaque, complex, and hard for non-experts to understand“(Pasquale, 2015) It is hard to figure out what can we learn from algorithm when it came out in the first time, just like being overwhelmed when you first seeing a computer.
After the birth of TikTok, the most notable hot spot of TikTok that triggered many social media platforms to start putting effort into research is its humanized algorithms. It is equipped with one of the best recommendation engines all over the world. As a new user of Tiktok, you don’t need to search for anything that you intend to watch, what’s instead they would send you a video pool of various categories. Depend on the intangible objectives, like how long you stay on the video or hit a like, click-through rate and so on. The system would narrow the scope of the video categories gradually and provide more videos at a high-interested rate to users. According to the data of what is recommendation engine on Amplitude, “Amazon was one of the first major ecommerce companies to pioneer content-based filtering and filed a patent for their system as far back as 2001.”(Franklin, 2022) Beyond that, Tiktok is being created much later than Amazon. The dominant method of its recommendation engine is called hybrid filtering, which combines content-based filtering and collaborative filtering to achieve a relatively effective function. It not only contains the feature of making suggestions that appeal to users’ current interests in content-based filtering but also combines collaborative filtering of provide with related content like “people watch this also watched…” within a similar cohort.
What’s on the contrary?
Interacting with algorithms in daily life more and more, are we training it or it is training us? Noticeably, one of the eye-catching features of TikTok compared to other social media platforms is the implementation of vertical viewing which shifts our watching behaviour. I know the intention of building it is to adapt to mobile devices to make sure viewers feel convenient while swiping videos. If it’s still in horizontal view, it is complicated to browse and flipping your phone around is annoying even if it is just a casual action. And that’s the main culprit of feeling laziness and fast delivery speed of interested content. Among the community of elder group, they see Tiktok as a time-consuming killer that it is easy to drag people into the content environment and forget about the regular routine. “5 mins in TikTok equals 1 hour in real life” – a common phenomenon of heavy TikTok users.
Case Analysis
TikTok fined £12.7m for misusing children’s data
Recently in April.5th 2023, BBC published a news edited by Shiona McCallum, Tom Gerken & Zoe Kleinman of TikTok has been fined £12.7m by the UK’s data watchdog for failing to protect the privacy of children. It estimated TikTok allowed up to 1.4 million UK children aged under 13 to use the platform in 2020. “The ICO( Information commission office) told BBC News TikTok had “taken no steps” to obtain parental consent.” (McCallum et al., 2023) This could be controversial, actually Tiktok is developing a feature called teenager mode which allows younger groups to receive suitable content and be able to set a time slot while using the app. Although the invention of this new mode has improved the overall environment of children’s content to some extent, it cannot be optimized fundamentally. Which is being said, the teenager mode did help in terms of parenting but in most cases, it is the most playful stage as children under 18 and these “TikTots”(children who have social media accounts, with TikTok being the most popular, despite being under the minimum age.) would take whatever it costs to break that rule. They are most likely to take advantage of shortcomings under the regulation. Additionally, Tiktok raises a sense of anxiety among parents about their children spending overtime on the platform, as they all know it is easy to create a sense of dependency on passively receiving information, further forming the behaviour of laziness at a younger age. In 2019, The average user spending time on TikTok daily is around an hour which could be an issue, as of now, Tiktok’s algorithm is developing so fast and comprehensively to feed various users. The spent time on it now is largely like to increase as time passes by. “TikTok is a hugely popular social media platform that has helped children keep in touch with their friends during an incredibly difficult year. However, behind the fun songs, dance challenges and lip-sync trends lie something far more sinister.”(BBC, 2021)
TikTok’s Datafication: A second thought incurs profound fear
Due to the sophisticated datafication of Tiktok, it has been called a data collection service that is thinly veiled as a social network. Your personal information including locations, cameras, records and so on will be shared once you log in to your Tiktok account. This is inevitable as it is asking you to complete a satisfaction survey, otherwise you cannot fully engage with all the features of Tiktok. It claims to do so to improve the overall user experience but you can not guarantee whether it is being used in other areas. This dramatically raises the question of seeing the person’s TikTok account you might familiar with in real-time as you are browsing videos because it is asked to allow access to your contacts, even recommend some accounts you don‘t know but might be interested in. We wonder why are we seeing this content from other users. This brings up the disputed thinking of the TikTok beauty algorithm, Referring to a blog on TikTok algorithm, “The Intercept (Biddle et al., 2020) leaked a document from TikTok in which there were instructions to its moderators to suppress any content from users who had ‘ugly facial looks’ or ‘abnormal body shape’ (including being ‘chubby, obese or too thin’) (Maragir, 2021) A Tiktok user Benthamite indicated Tiktok’s potential analysis of evaluating facial beauty. The higher achievement of standard beauty like youth and clear skin, the more exposure the video can get which means more tweet volume to all users. Therefore, this could be a way of collecting information and analysing data on facial recognition in the corporate and institutional spheres but it also challenges the social issues of appearance anxiety, which could cause panic among users about meeting the mass beauty standard. To be more specific, the beauty algorithm recognition actually creates a sense of techlash. It constrains the future development of AI technology and goes again the objective mass demand for stream media. This means that it should be categorized and judged by the content quality of the video and the degree of meaningfulness, instead of exterior elements like appearance, these are really the least priority in consideration. Furthermore, it would also lead to a violation of portrait rights in the commercial aspect. Some businesses or individuals may use a high standard of beauty appearing in videos for other purposes like propaganda without telling the person.
Conclusion
To sum up, in order to maintain the balance governance of AI and humans, we must take AI into account from multiple perspectives and pay explicit attention to the possible hazards AI may result in. Just like the meticulous algorithm of Tiktok is like a double-edged sword, they not only collect your information and serve you with the most optimal personalization content but also categorized you and divide humans into social classes in the broad database. Further creates an atmosphere of negative thoughts and shapes the public view. On a broader level, AI is the second form of human consciousness because it carries our thoughts and is meant to design for us to use. As a result, the point of AI is the process of interaction between humans and machines, interactivity gives artificial intelligence the life it deserves. As long as without the proactive interaction of humans, what’s left is just a pile of computers and wires.
References:
Flew, T. (2021). Regulating Platforms. Cambridge: Polity.
Crawford, K. (2021). Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. New Haven: Yale University Press. https://doi-org.ezproxy.library.sydney.edu.au/10.12987/9780300252392
Wang, C. (2020, June 7). Why TikTok made its user so obsessive? The AI algorithm that got you hooked. Medium. https://towardsdatascience.com/why-tiktok-made-its-user-so-obsessive-the-ai-algorithm-that-got-you-hooked-7895bb1ab423
Haoran. (2020, April 19). Tik Tok: Using AI to take over the world. Digital Innovation and Transformation. https://d3.harvard.edu/platform-digit/submission/tik-tok-using-ai-to-take-over-the-world/
Franklin, N. (2022, February 20). What is a recommendation engine? How recommenders work. Amplitude. https://amplitude.com/blog/recommendation-engine
McCallum, S., Gerken, T., & Kleinman, Z. (2023, April 4). TikTok fined £12.7m for misusing children’s data. BBC News. https://www.bbc.com/news/uk-65175902
Jacobson, D. (2023, March 23). Should the US ban TikTok? Can it? A cybersecurity expert explains the risks the app poses and the challenges to blocking it. The Conversation. https://theconversation.com/should-the-us-ban-tiktok-can-it-a-cybersecurity-expert-explains-the-risks-the-app-poses-and-the-challenges-to-blocking-it-202300
Maragir. (2021, January 11). TikTok’s algorithm and beauty datafication. Mar’s Media blog. https://themediastudent411501139.wordpress.com/2021/01/11/tiktoks-algorithm-and-beauty-datafication/
Eliaçık, E. (2022, May 9). AI’s invisible hand on daily life. Dataconomy. https://dataconomy.com/2022/05/artificial-intelligence-in-everyday-life/
Tiktok Xiaoke. (2023, March 1). 海外抖音为什么可以这么火_TikTok_全球_运营. 手机搜狐网. https://m.sohu.com/a/647882042_121669264
Kohen, N. (2022, December 12). Council post: How your business can take advantage of TikTok’s success with AI and ML. Forbes. https://www.forbes.com/sites/forbestechcouncil/2022/12/09/how-your-business-can-take-advantage-of-tiktoks-success-with-ai-and-ml/
TikTok sued for billions over use of children’s data. (2021, April 21). BBC News. https://www.bbc.com/news/technology-56815480
Be the first to comment