I will never forget the day I used ChatGPT for the first time, shortly after its release in late November 2022. I asked it a seemingly simple question: ‘What date was 900 days ago?’ As a new tool, it was reasonable to expect it to handle such a basic query. However, it fumbled the answer so badly that I thought this technology was a failure, and Google had nothing to worry about. Ironically, I ended up turning to Google to ask the same question, and it provided a perfect answer.
Since relying on ChatGPT and other generative AI tools to expedite various daily tasks has become second nature, I often wonder how we ever managed without them. Two weeks ago, I was ready to create a new Index ETF Portfolio. Given the vast array of options and choices, I turned to ChatGPT, Microsoft Copilot, and Google Gemini to help me narrow down my options and land on 2-3 ETF stocks that met my criteria. First, I would like to clarify that I did my own research first prior to resorting to these tools. In other words, using Google Search, Reddit forums, and other online research tools, I came up with a list of possible ETFs to look into. I then summarized each according to its key characteristics like past performance, MER fee, holdings, yield, and exposure to make it easier to narrow things down. Once I had all this data, I was ready for the next step: feeding it into Generative AI tools to summarize and help me make a better decision.
Generative AI Meta Analysis
I took all this data and fed it into 4 different generative AI tools, including the ones mentioned above. I phrased my query to leverage their analytical capabilities. Essentially, I wanted them to analyze my shortlisted ETFs based on the criteria I established, offering a fresh perspective to validate my choices or perhaps even uncover hidden strengths in lesser-known options. The analysis from the AI tools confirmed my initial top contenders, solidifying my confidence in their suitability for my portfolio.
Next, after each AI tool gave its response, I gathered and combined all the insights. I then fed this combined data set back into each system for one final analysis. This process aimed to identify and weed out any potential errors or inconsistencies (AI hallucination) in the initial responses. While it doesn’t guarantee complete elimination of issues, it served as a way to rigorously test the information and enhance my confidence in the final selections.
Launching a space mission might necessitate this level of rigor, but for my investments, it’s not a mission-critical situation. However, I believe the extra time spent upfront is a worthwhile investment. By putting in the legwork now, using various AI tools and testing different strategies, I can build a portfolio I’m truly confident in. This allows me to transition to a more passive management style in the future, with just occasional monitoring and adjustments.