What can artificial intelligence not do?

In recent months, there have been a lot of discussions, articles, and podcasts about what AI can do, but I wanted to ask a different question: what can't AI do?

Needless to say, some will be excited to say: there's nothing it can't do. Artificial intelligence is the most disruptive and transformative advancement in technology. It is and will continue to change the world and how we interact with digital spaces.

The author of this article believes that advocating a career in engineering is a bad AI hype trend

Some research reports have predicted a highly AI-oriented future for us, and these reports are all supported by statistical data.

• In 2018, the McKinsey Global Institute released a research report on the future of work, predicting that by 2030, the jobs of 400 million people around the world may be replaced by various artificial intelligence systems, tools and platforms.
• In March 2023, Goldman Sachs released a report on the global economic impact: generative artificial intelligence could replace a quarter of current jobs.
• Earlier last month, the World Economic Forum's Future of Jobs report echoed Goldman Sachs' predictions and went further by sharing specific data on the impact of artificial intelligence on jobs. Its estimates of the job outlook are chilling: 83 million jobs will disappear, while only 69 million will be created, a gap of 14 million.

Therefore, people have reason to fear that their jobs will be replaced in the future, or that the remuneration of existing jobs will not guarantee that the standard of living will not decline. People fear they will be ruthlessly weeded out, losing careers, jobs, and the means to support them in living the lives they want. There is even a list of jobs that may be eliminated in the future (mainly in mainstream high-paying jobs such as technology, media, and law), and the emergence of artificial intelligence will cause these jobs to disappear first. What's annoying is that the reports above don't offer any solutions or advice on what people should do to keep their jobs.

The author tweeted on May 2, 2023 that the output of artificial intelligence is not worth promoting

Instead of worrying about how artificial intelligence will take over our lives, it is better to hone early on the things that artificial intelligence cannot replace, such as those actions, tasks and skills that cannot be digitized or automated, because they all require continuous human beings. decision.

So I think there are three categories of human-driven decision-making capabilities that every sector and industry needs today and will continue to need for the foreseeable future.

**1. Contextual Awareness. **Referring to the background, it's not some sort of singular existence. There is more than one (and possibly more) types of context to consider in human society: cultural, economic, emotional, historical, place, political, situational, and social. And there's often a thread running through these different backgrounds. Economic perspectives are often not independent of social and/or political perspectives. No one in real life can extract these backgrounds as independent and equally distributed elements; ** no matter how many well-known artificial intelligence researchers try to construct some representative None of the mathematical equations of nature can achieve this goal. **

For example, Internet pictures like the one above must be analyzed from the perspectives of human history, location, and political consciousness in order to know that these pictures are deliberately fabricated to arouse a certain reaction from the public.

**2. Resolve conflicts. ****AI is not equipped to resolve disputes. **In fact, it came out rather unflattering. In 2016, Tay, an AI chatbot, was shut down within a day of tweeting out racist, misogynistic and other views. Looking at what happened recently in March 2023, the AI chatbot connected to ChatGPT in Snapchat gave inappropriate advice to an adult, saying that they were minors.

AI doesn't know when to shut up or not answer. Artificial intelligence is programmed to respond. But the quality, appropriateness and effectiveness of its responses remain questionable. Dealing with problems with complex contexts is an obvious area of weakness for AI systems, tools, and platforms. Artificial intelligence may provide humans with a basis for choice, but it cannot really help us make informed decisions.

**3. Critical thinking. ****The human ability to assess and evaluate our environment is one of our distinctive traits. Critical thinking is a combination of problem solving, curiosity, creativity, reasoning, and strategy. AI currently does not perform well at human levels of any of these elements of critical thinking. **

For example, computer programming, aka coding, does solve some math problems. However, as soon as any element in the mathematical equation represents a human being, then the solution to the problem through coding will not be the optimal solution, because at least one demographic group will be excluded and oppressed by it.

Another example, is so-called AI art creative? Some argue it's original, but it does have some serious copyright infringement disputes, raising concerns about the legality (and ethics) of content searched from online repositories.

The development of artificial intelligence has clearly reached a firm ceiling, because reasoning and judgment are carried out by balancing life experience, professional knowledge and skills. Artificial intelligence itself has no life experience, and its professional knowledge is also determined by certain training/test data sets and programmable skills, and its scope is limited. What is lacking still requires people to make up through critical thinking skills.

AI does not have situational awareness, conflict resolution, or the ability to think critically. If you want to pursue a career or have a skill that requires a combination of these three abilities, then you cannot be replaced by artificial intelligence. What is most likely to happen is that your job responsibilities will shift rather than disappear, requiring you to analyze the influence of multiple contexts, be able to manage tension, and think critically.

For example, a more critical ability for a programmer, whether he has a computer science degree or not, is to be able to explain and explain how generative artificial intelligence actually works. You can't be a good software developer without being aware of the potential social impact of the few lines of code you write. In essence, understanding the subtleties of our strengths and weaknesses in digital and communication skills in key domains of knowledge will become critical.