Why not use chatbots to predict where the price of Bitcoin will be by the end of 2025?
It is uncertain how valuable these are as predictive mechanisms. AI does offer the opportunity to predict future outcomes due to the fact that computers excel at pattern recognition. Since most things end up cyclical, this is a trait that can provide accuracy when looking into the future.
After all, history doesnt repeat itself but it sure does rhyme.
Part of this is due to the fact that human behavior is slow to change. In other words, we keep doing the same thing. It is why markets have an element of predictability, since they are driven mostly by sentiment.
So what does this mean for Bitcoin?
The Chatbots were tasked with generating outcomes. Let's see what they said.
ChatGPT 4o
Given the data, I project that bitcoin ( BTC) will close at $128,000 on Dec. 31,
ChatGPT o3
Base case: $127,500
Grok3
$105K
Deepseek “Deepthink R1”
Projected Closing Price of Bitcoin on Dec. 31, 2025: $138,000
Pi AI
$95,000
Mistral AI’s Le Chat
$120,000
ChatGPT o4 Mini
I project bitcoin will close at $115,200 on Dec. 31, 2025.
It is evident that ChatGPT is the most optimistic overall. It is only outpaced by Deepseek. Grok and Pi are fairly pessimistic, although Grok3 does see price appreciation going into the end of the year.
We have to keep in mind these are chatbot, trained for language. That is very valuable yet does not make it an "ai system". The industry is trying to add "thinking" and "reasoning" capabilities. However, without the vast array of data, it could be problematic.
Markets are moved by incalculable numbers of factors, since sentiment is driven by this. Anything has the ability to have an impact.
Data The Driver
As always, data is crucial.
For AI forecasts, the amount of data require is enormous. This is especially true when asking general questions for predictions.
At this point, most systems will be inadequate. We are probably better off looking at the betting markets since that has players who are driven by sentiment. This means that, if enough people are places bets, we could see a statistical reflection in the betting markets that mirror the larger ones.
Until infrastructure is built that can allow for mass data aggregation, it is likely that forecasts such as what these AI are doing is limited.
Nevertheless, it is fun to test them out.