Survival guide for the age of AI
Knowing the enemy’s weaknesses and devising our career strategy — 5 things that AI isn’t good at and 9 action plans for us
Until the early 20th century, following the Industrial Revolution, a profession known as the ‘Knocker-up’ existed in Europe. Their duty was to wake up people who needed to go to the factory early in the morning. Knocker-ups used long sticks to knock on customers’ doors or shot beans at their windows using a pea-shooter to create loud noises and wake them up.
Starting in the mid-20th century, when alarm clocks became commonplace, the need for people to knock on customers’ doors to wake them up naturally disappeared, and this job became obsolete. Countless jobs have been threatened by the invention of new technology, and in modern times, where technological change is even more rapid, more and more jobs are disappearing while new ones emerge.
In the past, creating subtitles and adjusting colors on film required manual labor after shooting a movie on film. However, after the advent of digital filming, jobs such as ‘movie subtitle creator’ and ‘film color correction specialist’ became obsolete. Similarly, the role of the ‘flight engineer,’ who assisted pilots in checking and maintaining the proper functioning of engines and equipment, has been eliminated due to the automation of navigation devices.¹
Recently, AI has demonstrated great potential in fields such as writing, coding, composing music, and creating art, posing a rapid threat to jobs that were previously believed to require human ‘creativity.’ The performance of AI is expected to improve even more rapidly in the future, as they utilize rich data collected from AI applications and reduce the cost of computing and machine learning.
In the midst of this trend, it seems unlikely to win a head-on battle against AI. Sequoia Capital has predicted that by 2030, AI will be able to produce final drafts of writing that surpass those of professional writers, write better code than full-time developers, and create better images than professional artists, designers, and photographers.²
However, upon examining the quality of results produced by current AI tools such as ChatGPT and Midjourney, it becomes clear that AI still has limitations in its ability to completely replace humans at present. To increase our chances of survival, we need to develop skills that complement AI and allow us to coexist with it. We need to excel in areas where AIs are not as proficient.
Let’s explore 5 weaknesses of AI that we can exploit.
1. Out-of-the-box thinking
First, we need to understand how AIs work so that we can grasp their strengths and weaknesses. The following image shows the operation of large language models (LLMs) such as ChatGPT.³
The first step is to sample a prompt and have a human labeler write a response. Next, a labeler evaluates several options presented as responses to the prompt and ranks them from best to worst. This data is then used to train the AI model to generate a response to the prompt, and the process is repeated, using the resulting data to improve the model’s learning.
Large language models identify statistical regularities in text. Any analysis of the text of the Web will reveal that phrases like “supply is low” often appear in close proximity to phrases like “prices rise.”
The fact that ChatGPT rephrases material from the Web instead of quoting it word for word makes it seem like a student expressing ideas in her own words, rather than simply regurgitating what she’s read; it creates the illusion that ChatGPT understands the material.
- Ted Chiang⁴
The words that ChatGPT produces are combinations of the most statistically likely expressions. This is precisely why we feel that ChatGPT is smart. Ted Chiang argues that this is just an ‘illusion.’ ChatGPT’s ability to insert ‘prices rise’ after ‘supply is low’ is due to its learning from input data that has been ranked as the most plausible expression. It is not the result of inferring economic logic.
Thus, while AI that relies on large amounts of input data may excel at combining statistically plausible results, it lacks the ability to come up with unconventional ideas or original hypotheses. It can easily put together 100 to make 101, but it cannot create zero to one. Creating something unique from scratch, with out-of-the-box thinking, is not a strength of AI.
2. Messy and complicated communication
To obtain accurate results from AI, we need to provide clear prompts. However, human communication is often incomplete and ambiguous. It is rare for someone to know exactly what they want, and it is not easy to convey our thoughts precisely as intended.
We often have to supplement missing context ourselves to understand others. For example, when your boss suddenly asks, ‘What happened with that thing?’ we need to connect the dots in our mind, considering the time, place, and occasion in order to infer their intent. Sometimes we require information that only insiders know, or knowledge of social and cultural backgrounds to truly comprehend others’ messages. We may need to ask follow-up questions to clarify any ambiguity.
As the number of conversation participants increases, the complexity of communication explodes. Two people need one channel to communicate. For three people to communicate, they need three channels: A to B, B to C, and C to A. For four people, six channels are needed. So, what if you have to collaborate with 20 teammates and communicate with 30 clients? Can AI handle the myriad complexities and ambiguities that arise in this process?
For the time being, AI won’t be able to take the lead in creating new insights, solving problems, and devising strategies by aggregating information obtained through complicated communication. If the world were neatly arranged like a sewing room, it might be a different story, but in reality, it’s more like a warehouse full of tangled balls of thread. It’s a difficult task for AI to untangle the threads and weave a new garment.
3. Emotional resonance
We need various soft skills in the workplace. If you are a leader, simply being good at your job does not necessarily mean you can lead a team effectively. You often need to trust your intuition to make decisions and show qualities such as passion and perseverance. Leaders should also know how to be persuasive and provide proper encouragement. These are the qualities that machines cannot possess.
Empathy, the ability to put oneself in another’s shoes, is crucial for human work. To build and pursue organizational goals, it is essential to actively listen to stakeholders including teammates, clients, and customers and understand their perspectives. The ability to simulate what other people think and feel is a unique human ability. AI may appear to empathize, although it is only an imitation and cannot truly demonstrate cognitive or emotional empathy.
4. Physical labor
In the past, many people believed that robots would replace simple physical labor, and that creative professions would survive. With technology evolving, robots that can walk on two legs, ride bicycles, run, or jump have been developed and have shown great potential.
However, it is becoming increasingly clear that they still have a long way to go before they can fully replace human physical abilities. Jobs that require not only exceptional physical skills, but also complex communication skills, such as experienced technicians, are unlikely to be replaced by AI anytime soon.
5. Critical thinking
ChatGPT exhibits something like the banality of evil: plagiarism and apathy and obviation.
- Noam Chomsky⁵
When you ask ‘What is your political viewpoint?’ to ChatGPT, you get this answer.
As an AI language model, I do not have personal beliefs or viewpoints, including political ones. My function is to provide informative and helpful responses to users’ questions using the information that I have been trained on. I am designed to be neutral and to provide a broad range of perspectives and information without taking any positions or promoting any particular agenda.
When asked about its position on a controversial issue, ChatGPT avoids it by saying it ‘does not have a viewpoint.’ This is a limitation of AI, which merely follows received commands and cannot express its own subjective opinion. As it generates responses based on data collected from the web, it cannot distinguish between truth and falsehood. Additionally, it cannot make ethical judgments. It is difficult to say that AI has true intelligence when it lacks its own perspective.
Hannah Arendt said that ordinary and mundane people can perform evil. She argued that atrocities in history, such as the Holocaust, were not committed by fanatics or sociopaths with evil intentions, but by people who were thoughtless and simply stood by. Like the saying ‘Silence is complicity.’ Noam Chomsky suggests that AI chatbots like ChatGPT exhibit a similar concept, the ‘banality of evil.’⁶
Having one’s own perspective is not easy. It requires gathering, digesting, and analyzing a vast amount of information. Only those who have trained themselves to do so can draw conclusions, make decisions, and fight for their beliefs. AI’s indecisiveness displays a lack of critical thinking, which is a significant difference between humans and machines.
Now that we’ve discussed the weaknesses of artificial intelligence, let’s take a look at 9 action plans that can increase our chances of survival.
If you know the enemy and know yourself, you need not fear the result of a hundred battles.
— Sun Tzu
1. Retire
I am not kidding. The most foolproof solution is to retire early. The rise of AI is causing wealth inequality, which means that as companies and individuals who use AI to amass great wealth continue to grow, many people will lose their jobs. It might be a practical idea to aim for financial freedom before such a world arrives.
I know, it’s easier said than done. If achieving complete financial freedom seems difficult, consider generating passive income partly. You can study and select companies that will benefit greatly from AI and invest in them. It will be like you and AI making money together. And don’t worry, you can start small.
2. Do physical labor
In the last century, sci-fi movies often depicted robots as a regular feature. People believed that robots would replace human labor in the future. With factories being automated and microcomputers being embedded in almost all household appliances, robots have certainly changed the world.
However, it has been revealed that it is quite difficult for robots to replace all human physical activities. While we now live in a world where robots in warehouses can pick up and dispatch items automatically, the last mile delivery is still a challenging task even for the smartest robots. There will always be areas that require human physical labor.
Jobs that require both complex communication skills and physical abilities are difficult for robots to perform. Examples of such jobs include technical roles such as construction workers, mechanics, and maintenance workers, as well as medical-related and emergency response jobs like physical therapists, medical technicians, firefighters, police officers, and soldiers.
There was a 1987 film called ‘RoboCop’ that depicts the story of a police officer whose body is replaced by a robot. As in the movie, the use of robotics to enhance and assist human physical abilities will become more common. However, will robots be able to perform police duties independently, completely replacing humans? That would require a considerable amount of time.
3. Be unexpectedly creative
We do not seek highly probable theories but explanations; that is to say, powerful and highly improbable theories.
— Karl Popper
We learned how ChatGPT can generate the most ‘statistically plausible’ combinations. That’s why AIs can produce text that sounds ‘believable’ very quickly and easily. If we want to stand out and surpass what AI can do, we need to bring unique human insights to the table.
Similar to AI, humans also need input to generate ideas. But AI’s input mostly comes from the web. We humans need to approach this differently. The importance of reading and studying in various fields is still relevant. We should also engage in activities that only humans can do, such as enjoying nature while taking a walk or savoring good food at a restaurant, fully utilizing our senses.
The process of humans creating ideas is completely different from that of AI. We have the ability to connect the dots in our heads and have that ‘aha!’ moment. The advertising legend James Webb Young suggested that we can deliberately create ideas in his book ‘A Technique for Producing Ideas’, by following steps of gathering raw material, digesting it mentally, and processing it unconsciously.
The capacity to bring old elements into new combinations, depends largely on the ability to see relationships.
— James Webb Young⁷
Unlike AI, which is designed to generate ‘believable’ results, humans can utilize various areas of the brain to come up with creative hypotheses or conclusions that go beyond the norm. If we have a unique perspective or the ability to see relationships, we can create exceptional ideas that are qualitatively different from the plain results produced by AI.
4. Manage complexity
Artificial intelligence excels at producing predictable outputs from simple inputs. Consequently, many people predict that numerous entry-level office jobs, which require little collaboration and involve relatively straightforward tasks, will soon be replaced by AI.
However, jobs that involve communicating with various stakeholders and processing complex information will require human abilities. Tasks such as hosting meetings, reconciling different opinions, mentoring team members, making difficult decisions, developing strategies to solve complex problems, and managing constantly changing environments are too difficult for AI.
To tackle complexity, one must accumulate experience and refine skills over time by starting with simple tasks. In old martial arts movies, masters often assign their students menial tasks like cleaning the floor before teaching them martial arts. There is a reason for it. In my own experience, I have typed countless meeting minutes and created wireframes like a machine, which paved the way for me to take on more challenging roles.
AI has the potential to erode entry-level jobs, potentially limiting opportunities for career growth into higher-level positions that require more complex skills. It is therefore critical to accurately assess the skills needed to progress in your career and develop an effective learning plan. Blindly delegating tasks to AI for convenience can hinder your professional development. It is crucial to learn how to discern when to utilize AI and when to stay away.
5. Exercise leadership
While AI may quickly take over many entry-level tasks, replacing leaders will be a much greater challenge. Leaders need to establish the organization’s vision and guide team members towards achieving the goals. Merely being skilled at working independently or having a lot of knowledge is not enough.
Leaders who possess qualities that machines can’t replicate, such as intuition, decisiveness, perseverance, a sense of responsibility, strategic thinking, and strong interpersonal skills, will continue to be in demand as valuable assets in many workplaces.
6. Empathize
Actively listening to other people and sincerely understanding the background and situation is a unique human ability. It won’t be so easy to replace professions that require empathy, such as counselors, psychotherapists, educators, and social workers, with machines.
While various robots are being introduced in the medical field, jobs such as doctors, nurses, and physical therapists still require the ability to understand and care for others’ emotions. Professions that involve listening to concerns and solving problems, such as lawyers, consultants, and crisis management specialists, will also continue to be in demand.
7. Have your opinion
If you live a life where you are swayed by others’ words and make important life decisions based on them, it’s not much different from what AI does — producing the most statistically plausible outcome.
We’ve long called people who mindlessly repeat others’ words ‘parrots’, but lately I’ve noticed how much they resemble ChatGPT. That’s why I suggest adding this term to the dictionary.
—
ChatGPT
[tʃæt dʒipiti] ・noun
A metaphor for people who mindlessly copy what others say without engaging in critical thinking or independent analysis.
“It is creepy how he just parroted Mary’s words and presented them as his own original idea. He’s such a ChatGPT.”
—
Although it’s important to be open to others’ opinions, it’s equally crucial to find your own voice and stand firm in your beliefs. If you have ideals you want to achieve and values you want to uphold, it’s important to demonstrate them through your words and actions to impact the world.
Only those who think for themselves can take control of their own lives. Live YOUR life, not the one that seems most statistically likely.
8. Combine the above qualities
Combining the measures presented above will become an even more powerful weapon to resist AI.
- A Skilled technician who ① Exercises physical labor and ④ Manages complexity
- A strategist who ④ Handles complexity and ③ Thinks Out-of-the-box
- A manager who has ⑥ Strong empathy and can ⑤ Exercise leadership
- An artist who ⑦ Has an unique perspective and can ③ Create innovative ideas
- An healthcare worker who ① Does physical labor and ⑥ Exhibits empathy
Envision a career that fits you and think about what skills and qualities you need to develop. If you can combine multiple aspects that machine can’t replicate, you don’t need to concern about losing your job to AI and can become a rare talent in the job market.
9. Be persevere
Survival can be summed up in three words — Never give up. That’s the heart of it really. Just keep trying.
— Bear Grylls
From the time we start working until retirement, it usually takes around 40 years, and the career path is never a smooth upward trajectory. There may be tough times when economic crises hit or jobs disappear due to changes in the industry. But no matter what challenges come your way, if you don’t give up, there will always be a way to survive and thrive.
I hope this article helps you build a great career and I wish you the best of luck. Thank you for taking the time to read.
I wrote this article, inspired by Mirum Hong Kong UX team’s learning session about artificial intelligence. Thank you team💜
References
[1] What are the 18 recently disappeared occupations that gradually decreased to zero?, jobsN
[2] Generative AI: A Creative New World, Sequoia Capital
[3] Training language models to follow instructions with human feedback, arXiv
[4] ChatGPT Is a Blurry JPEG of the Web, New Yorker
[5] Noam Chomsky: The False Promise of ChatGPT, The New York Times
[6] Eichmann in Jerusalem: A Report on the Banality of Evil, Hannah Arendt
[7] A Technique for Producing Ideas, James Webb Young
Articles written around this topic
