Summary of Key Points
Recently, AI has frequently caused problems in various fields such as public service advertisements, award-winning competition entries, and everyday practical applications (e.g., identifying mushrooms, checking flight tickets, and writing research reports). However, most users immediately blame AI for these issues. The article argues that this approach is unreasonable: AI is merely a tool, and the ultimate responsibility lies with humans. Over-reliance on AI stems from a lack of understanding of its technical principles. To make effective use of AI, one must possess the "tool literacy" required to understand its limitations, know its capabilities, and be able to verify the results. At the same time, both AI providers and regulatory authorities must also fulfill their responsibilities to promote the proper use of this technology.
Detailed Analysis
1. AI's Various Mistakes, but the Blame Always Falls on It?
There have been numerous amusing and problematic incidents involving AI recently: in a public service advertisement in Linfen, Shanxi, a pedestrian was depicted with their head and feet reversed; in a photography competition in Hohhot, the first prize-winning image was generated by AI; in Dongguan, an AI-generated calligraphy work won an emergency书法 award; people have lost money using AI to identify poisonous wild mushrooms, made mistakes when checking and changing flight tickets, and even experienced disappointment when booking banquet rooms. In one legal case, a lawyer used an AI-generated argument that was refuted by the judge in court. In almost all of these situations, the blame was placed on AI, making it the scapegoat.
2. Is It Reasonable to Blame AI?
Common sense and the law both disagree with this notion. AI is simply a tool, similar to a kitchen knife or a road. A kitchen knife can be used to prepare food or to cause injury, and a road can be used for walking or leading to accidents; the key lies in how it is used. The output of AI is determined by human instructions, and it can also generate incorrect or misleading information (a phenomenon known as "AI hallucination"). If you don't verify the results after using AI, blaming it for errors is like blaming an uneven road or a sharp knife for an accident.
Legally, AI is not considered a "person" and does not have civil rights or responsibilities. For example, if an employee makes a mistake while using AI to draft a contract, the company can hold the employee accountable; if a student plagiarizes using AI, the school can handle it as academic misconduct; if a public service advertisement fails to meet quality standards, the relevant departments must be held responsible. The ultimate responsibility always lies with the person using the technology.
3. Why Do People Over-Trust AI?
Many people trust AI blindly because they do not understand how it works. AI does not actually think; it learns patterns from large amounts of data and then provides "most likely" answers based on probability. For instance, when asked a question, it may provide an answer that seems correct but is actually fabricated (an example of "AI hallucination"). Since its answers are based on guesses, one cannot rely on it completely, as blind trust can lead to mistakes.
4. To Use AI Effectively, One Needs "Tool Literacy"
While AI doesn't require a "license," using it properly requires three key skills: 1) understanding its underlying principles (i.e., that it makes probabilistic predictions rather than real conclusions); 2) being aware of its limitations (e.g., not using it to write entire research reports or make medical diagnoses); 3) verifying the results (e.g., checking the information obtained from AI before accepting it as fact). It's fine for ordinary people to use AI to check the weather or find restaurants, but it would be risky for analysts to rely solely on AI to write research reports or for doctors to rely on it for medical diagnoses. Without proper understanding, users could become victims of fraud.
5. AI Providers and Regulators Must Also Be Responsible
In addition to users, AI companies must also take responsibility: they need to improve their products to reduce errors, disclose the parameters and risks of their technology, and provide education to help people better understand how AI works. Legal and regulatory authorities should establish clear rules and enforce them (e.g., by penalizing companies that fail to meet quality standards). Only by doing so can the AI industry develop in a healthy manner and users can use it with confidence.
Conclusion
AI is a valuable tool, but it should not bear the blame for all mistakes. Everyone using AI must take responsibility and improve their understanding of the technology. Both providers and regulators must play their roles to ensure that we can fully benefit from its capabilities.