Summary of the Key Points
This article focuses on the “AI for Good” global summit hosted by the United Nations. It discusses AI projects that address real-life challenges faced by ordinary people, such as the “Field Sentinels” in Kenya and the endangered language archives in Tanzania. However, it also raises three critical questions about the concept of “AI for Good”: whether these projects can be sustained, whether they are truly necessary, and whether the benefits are fairly distributed. The article points out that AI has already permeated our daily lives (for example, through price discrimination based on user data and algorithms that exploit delivery riders). Yet, ordinary people are not powerless; by taking specific actions—such as turning off personalized recommendations, filing complaints, and demanding explanations for decisions—the impact of AI can be mitigated. It emphasizes that “AI for Good” is not an inherent aspect of technology but requires the participation of everyone.
Detailed Analysis
1. AI at the Summit: Solving Real Problems, Not Just Showoff
The AI projects presented at the summit were practical solutions to common issues faced by people:
- Field Sentinels: Engineers in Kenya created solar-powered smart camera poles that cost only 20 Kenyan shillings per month and issue alerts within 5 seconds of detecting pests or diseases, helping farmers reduce losses by 30%. This solution addresses the question of how the poorest can afford the latest technology.
- Endangered Language Archives: A Tanzanian team built a free database of recordings in languages with no digital records, as ChatGPT only understands a few dozen languages. They fear that these languages will disappear in the digital world.
- AI and Grief: Forty percent of young people use AI to handle the affairs of their deceased relatives or to talk to friends in the evenings. While technology has entered this private emotional space, there is little discussion about how it should be used.
There was also a space competition in space computing, where “feasibility” accounted for 40% of the scoring criteria. Winning teams from Asia, Africa, and Latin America used satellites to monitor food and water quality. The common theme of these projects is that the focus was on the practical difficulties people face, not on the technology itself.
2. Three Critical Questions about “AI for Good”
The article raises three straightforward but important questions:
- Can these projects survive beyond the initial hype? There are many examples in the industry where projects are launched with great fanfare but later abandoned due to a lack of maintenance.
- Would things be cheaper without AI? In many places, what is lacking is not diagnostic software, but doctors, medicines, or emergency services. If using AI would make things cheaper, it might just be another form of “greenwashing” by companies.
- Who benefits from the technology? While AI giants have trillions in market value, behind them are data annotators in countries like Kenya and the Philippines who earn very low wages. They train AI to recognize objects or manage traffic signals but are rarely mentioned in promotional materials about “AI for Good.”
3. AI Is Already Making Decisions for Us, We Just Don’t Realize It
AI is influencing our lives every day:
- Price Discrimination: The same hotel charges more to regular customers (Ms. Hu from Shaoxing sued the platform and won a three-fold refund).
- Algorithms Exploiting Riders: Delivery times are being shortened because algorithms set a standard based on the fastest delivery times, even though a “wait 5 minutes” option was later added.
- Resume Screening: Job seekers in the U.S. were rejected after submitting 100 applications; changing their birth dates led to interview invitations (the company compensated 200 people).
- Foreign Examples: In the Netherlands, the tax system used AI to target fraud, causing 35,000 families to go bankrupt and leading to the resignation of the entire government. In the UK, AI was used to grade public school exams, which sparked student protests and led to a return to manual grading by teachers.
4. What Ordinary People Can Do
You don’t need to understand AI technology to protect your rights:
- Turn off Personalized Recommendations: Many apps have settings to turn off personalized recommendations; just search for the option (many people don’t know it exists).
- Compare Prices and File Complaints: Use family members’ phones to compare prices and screenshot the results to file complaints with relevant authorities (e.g., 12315 or the platform).
- Demand Explanations: When loans are denied, resumes are ignored, or claims are rejected, don’t accept vague system explanations.
- Warn the Elderly about Scams: Explain how AI can be used for fraud and help them stay vigilant.
- Speak Out About Issues: Many cases of price discrimination, improved delivery times, and changed exam grades were resolved because ordinary people spoke up.
Conclusion
“AI for Good” is not just a slogan; it’s a tug-of-war involving everyone. On one end are capital and computing power, and on the other, you and I. Next time you hear about “AI for Good,” ask yourself these four questions: Who does it benefit? Who makes the decisions? Who pays the cost? And who reaps the benefits? Only initiatives that can withstand such scrutiny are truly “for good.” If you don’t take action, the balance will tilt in favor of capital.