I came across a post on X about predictions for condominium prices. It referenced an article from the Weekly Gendai magazine, published on August 25, 2018. In the article, AI had predicted the percentage increase or decrease in second hand condominium prices seven years into the future—that is, in 2025. To summarize the content simply: “Condominium prices are expected to decline.”

Skipping the detailed contents, the results were way off. For example, take “Daikanyama Address The Tower” (in Shibuya, Tokyo). The prediction estimated a price change of -11.6% from 2018 to 2025, equating to a decrease of 88.4% of the original value. The forecasted price, adjusted to a 70-square-meter basis, was projected to fall from 118.3 million yen to 104.62 million yen—a drop of over 10 million yen. In reality, however, a currently listed sale price for a unit in the same building found online (reference: 238 million yen for 65.97 square meters) translates to approximately 252.54 million yen for 70 square meters. That’s a miss of about 150 million yen.

I don’t know how many people believed this article and went ahead with selling their property, but those who did must be feeling quite regretful. The AI’s prediction ended up being unfortunately inaccurate.

That said, I’m not trying to claim that “AI can’t be trusted.” Here are three thoughts I had after reading this “missed prediction” article:

1. AI is not all-powerful

The AI of 2018 probably wasn’t as advanced as today’s, and we don’t even know how the data in the article was processed by AI—if it truly was AI at all. However, even if we use the latest AI to make future predictions, there is still a significant risk of being way off. In the first place, AI is not particularly good at predicting the future.

What AI is good at includes generating responses based on training data, streamlining simple or repetitive tasks, and interpreting large sets of structured or unstructured data—such as extracting insights from big data. Features like creating meeting transcripts from voice recordings fall into this category.

But when it comes to future predictions, AI struggles. It performs well when forecasting equipment failures or consumer behavior based on massive datasets, but only when those predictions are based on stable trends from the past. AI is weak against discontinuous changes like pandemics, wars, or financial crises. Furthermore, real estate prices also reflect human factors—such as livability and environmental concerns—which are outside AI’s strengths. At present, making highly accurate real estate price forecasts is difficult for AI.

2. Don’t blindly trust mainstream media

This doesn’t apply only to real estate—TV and weekly magazine reporters are not necessarily experts in the fields they cover. In fact, they are often quite unfamiliar with the subject matter. The same goes for their readers. People who try to gain knowledge from TV shows or magazine features are usually not well-versed in the topic. That’s why rather than writing serious articles with core information, media tends to favor sensational and entertaining content which appeals to their audience. Naturally, the reliability of such content is questionable. To gain accurate knowledge and understand real trends, one should refer to economic or academic journals, or reports published by governments and corporations.

3. You are the one who predicts the future

When AI or mainstream media presents a “prediction,” it can feel like the correct answer. However, what truly matters is how you interpret and utilize that information—it’s about the mindset of the person receiving it.

AI excels at data analysis, but it struggles when it comes to futures influenced by society and human emotion. Even if predictions such as election results are accurately calculated, the mere spread of that prediction can alter people’s behavior and ultimately change the outcome. In this sense, the prediction itself can influence reality. Moreover, magazine predictions often prioritize sensationalism. The quality of information is inconsistent.

What’s essential in today’s world is the ability to evaluate information and have your own standards for judgment. Predictions are not something to be blindly believed—they are just one of many tools for making decisions.

AI is certainly convenient. It performs impressively when there’s already a correct answer. But when it comes to predicting the future, as mentioned earlier, there’s no way to know if its answer is truly correct. “Predictions” are seen everywhere, especially from AI which seems to produce plausible answers—but you must not take those outputs at face value. The futures described by AI or the media are merely one possible reference. Ultimately, the one who predicts and shapes the future is you.

For additional information or any questions please contact us here
Email: info@remax-apex.com