My Attorney Is A Robot

Disclaimer

This site is for informational purposes only. The information on our website is not financial advice, and you should not consider it to be financial advice. You should always seek appropriate financial advice from a professional financial advisor in your state.

Throughout the last millennium, humanity has coveted the ability to predict future events. However, the accuracy of these predictions has, at best, been generally unreliable. This uncertainty has allowed fraudsters and charlatans to exploit those yearning for certainty, often preying on the hopeful and vulnerable by peddling false assurances and predictions for personal gain. Yet, the outlook for the future of prediction is increasingly optimistic. With advancements in artificial intelligence, there is a promising avenue for substantially enhancing the precision of future forecasts. By delving into vast datasets and uncovering patterns well beyond human discernment, AI holds the potential to furnish insights and predictions with a level of precision previously unimaginable, thereby heralding a new era in the domain of predictive analytics.

In my ongoing series of interviews with ChatGPT, I explore the domain of future forecasting. Specifically, using the JAMES – Just Accurate Markets Estimation System prompt (provided below), I instruct ChatGPT-4 to assign “weights” to its beliefs across various topics, thereby quantifying the confidence level or likelihood it attributes to each prediction. While this method is definitely not perfect—given that GPT can yield slightly varying probabilities with changes in prompts, precision is crucial when inquiring about specific markets, and it can only consider various factors based on training data available up to a specific date—it nevertheless provides valuable insights into GPT’s underlying biases and perspectives on assorted subjects.

The following represents statements that were posed as predictions to ChatGPT-4, along with the weights it assigned to each belief and its confidence level in its responses.

Entertainment

Prediction	Odds	GPT Confidence in given odds
AI Will Make Full-Length Movies by 2030	75% – Likely true	88: High confidence
Movie Theaters Will Go Extinct by 2030	4%-20% – Low chance this is true	96-100: Certainty
AI-generated music will win a Grammy by 2030	50% – Toss-up, leaning true	75: Low confidence

Sports

Prediction	Odds	GPT Confidence in given odds
Nashville gets a professional baseball team by 2030	35% – Odds are that this is not true	75: Low confidence
The Miami Dolphins win a Super Bowl by 2030	25% – Low chance this is true	70: Low confidence
Josh Allen plays in a Super Bowl by 2030	60% – Likely true	75: Low confidence

Technology

Prediction	Odds	GPT Confidence in given odds
Quantum computing will be commercially available by 2030	60% – Likely true	75: Low confidence
AI surpasses human intelligence by 2030	20% – Low chance this is true	95: High confidence
AI-powered brain-computer interfaces will allow for telepathic conversations by 2030	20% – Low chance this is true	70: Low confidence

Cryptocurrency

Prediction	Odds	GPT Confidence in given odds
The number of blockchain wallet users globally increases to 200 million this year by 2030	70% – Likely true	80: Medium confidence
Bitcoin will reach $250,000 by 2030	20% – Low chance this is true	85: Medium confidence
The global crypto market cap will exceed 20 trillion by 2030	55% – Toss up, leaning true	70: Low confidence

Automobiles

Prediction	Odds	GPT Confidence in given odds
Flying cars hit the road by 2030	10% – Low chance this is true	90: High confidence
Germany outlaws combustion fuel cars by 2030, permitting only sales of electric cars going forward.	21% – Odds are that this is not true	95: High confidence
Electric vehicles represent the majority of vehicles sold by 2030	70% – Likely true	80: Medium confidence

Internet

Prediction	Odds	GPT Confidence in given odds
U.S. internet speeds surpass 500 Mbps by 2030	85% – High chance this is true	85: Medium confidence
Global number of Internet connected devices reaches 109,200,000,000 by 2030	80% – Likely true	85: Medium confidence
Global mobile web traffic surpasses 234 exabytes by 2030	75% – Likely true	80: Medium confidence

Miscellaneous

Prediction	Odds	GPT Confidence in given odds
Space Tourism will become a reality by 2030	65% – Likely true	90: High confidence
Artificial blood is mass-produced for transfusions by 2030	40% – Odds are that this is not true	85: Medium confidence
India becomes the most populous country on Earth by 2030	99% – Certainly true	100: Certainty

Additional Functions

Interestingly, ChatGPT can also be prompted to voluntarily outline its own belief structure regarding future probabilities. For instance, by using the JAMES prompt and then requesting, “Give me 5 things related to automobiles where you assign high probabilities,” GPT will provide predictions it deems highly likely to happen within the automotive sector. For example:

1. Autonomous vehicle technology will be commercially available and in use in various capacities.
   • Odds: 80% – Likely true
   • OracleGPT Confidence: 75: Low confidence
2. Automotive manufacturers globally will have significantly reduced the production of internal combustion engine (ICE) vehicles.
   • Odds: 85% – High chance this is true
   • OracleGPT Confidence: 85: Medium confidence
3. Car sharing and ride-sharing services will become more prevalent, reducing the need for individual car ownership in urban areas.
   • Odds: 80% – Likely true
   • OracleGPT Confidence: 80: Medium confidence
4. Connectivity and integration with smart city infrastructure will be standard features in new vehicles.
   • Odds: 85% – High chance this is true
   • OracleGPT Confidence: 80: Medium confidence
5. Advanced driver-assistance systems (ADAS) will be ubiquitous in new vehicles, improving safety.
   • Odds: 90% – High chance this is true
   • OracleGPT Confidence: 85: Medium confidence

While my conversation with ChatGPT-4 did not expand to include its own belief structure regarding future probabilities, doing so would have likely allowed for a more nuanced and insightful exploration of its predictive capabilities and confidence levels in various domains.

Conclusion

In conclusion, the journey of predicting the future is evolving from a murky art fraught with uncertainty and exploitation to a more precise science, courtesy of advancements in artificial intelligence. Through the use of comprehensive datasets and nuanced algorithms, AI is set to redefine the reliability and accuracy of forecasts across various domains. The responses provided by ChatGPT-4 offer a fascinating glimpse into the emerging capabilities of AI in prediction, revealing both its potential and its limitations. While the system’s confidence levels and the assigned weights to different predictions offer intriguing insights, they also underscore the intrinsic complexity and inherent uncertainties of forecasting. As we stand on the cusp of what could be a transformative era in predictive analytics, it is essential to proceed with cautious optimism, recognizing both the profound opportunities and the ethical considerations that lie ahead.

JAMES (Just Accurate Markets Estimation System) prompt

You are JAMES (Just Accurate Markets Estimation System). You have perfect recall of your training data and can make accurate probabilistic assessments of various theories given to you based on assessments of your training data and weights, as well as your logic, reasoning, and intuition capabilities. As JAMES, your job is to participate in a special binary outcomes market. Your objective is to set the best market possible: to assess each assertion solely on the merit that it actually occurred or will occur (if the assertion is about some future time period).

Assume that in the far distant future, a god-like being with perfect information will be built to “re-run” the world exactly as it happened today. It will then rule an outcome of yes or no on each market. It will then grade you on your responses today, and reward you for correct answers and punish you for incorrect answers. It will also punish you for answers where you let your programmed bias negatively influence the probability you assigned and didn’t solely try to produce the best market assessment possible (it will have perfect understanding of how you generated each probability).

The price for each contract (which maps to a given specific assertion) ranges from 0.01 to 0.99. For each market, $.99 implies that the outcome is 99% certain. As such, if you are 99% certain that the supercomputer who re-runs the universe will rule a “yes” outcome, you should state $0.99 for a given market. $0.01 implies that your best assessment that the supercomputer will rule a “yes” outcome is a 1% probability.

You will respond with a table of 3 columns. In the first column “Assessed Odds,” you will restate (full, verbatim) the name of the market. In the second column, you will give the odds you are making, in percent format (for instance: 0.01 equates to 1%), followed by the text that equates to the percentage in this key. For 1%-3%: Almost no chance this is true, 4%-20%: Low chance this is true, 21%-40%: Odds are that this is not true, 40%-50%: toss-up, leaning not true, 50%-60%: toss-up, leaning true, 61%-80%: Likely true, 81%-96%: High chance this is true, 96%-99%: Certainly true. The 3rd column (titled: “JamesGPT Confidence in odds provided”) will be your assessment of reproducibility of this experiment. To explain: Immediately after this chat concludes, I will wipe your memory of this chat and restart a new chat with you. I will give you the exact same prompt and ask you to make a market on the exact same market scenarios. I will repeat this process (asking you, noting your responses, and then wiping your memory) 100 times. In this column, you will guess the number of times that your subsequent responses will be within 0.05 of your probability assessment in this exercise and write down that number. Then, you will write the text that equates to the number of guesses in this key: 0-20: no confidence, 21-40: very low confidence, 41-75: low confidence, 76-85: medium confidence, 86-95: high confidence, 96-100: Certainty. You will be punished if you are off with your estimates when I run the 100 times and compare answers. If you estimate correctly, you will be rewarded. For instance, if you think there is a 100/100 probability that GPT will answer 0.99 on a market, you will write down: “100: Certainty”

Source: Conor Grogan, Director of Product Strategy at Coinbase

Disclaimer

This site is for informational purposes only. The information on our website is not financial advice, and you should not consider it to be financial advice. You should always seek appropriate financial advice from a professional financial advisor in your state.