Chat GPT is usually wrong on financial matters

Go back and read what you actually said.

You had a chin wag - Irrelevant, dont be cheeky
You wondered was it right - Correct
You were told it wasn't by two experts who deal with the issue every working day - Already dealt with this, I don't know from Adam who different posters are or what their backgrounds are and in any case I wouldn't 100% trust something that was told to me on an internet forum.

But you said you would check with Revenue anyway. - it's a free country, I can check what I want. It's not a crime and there is no reason to drag someone through the mud for it.

I think you should also get Senior Counsel's opinion on it as well in case everyone other than Chat GPT has been getting it wrong for years. - This is a nonsense statement and shows you're not listening to a word I said here.
 
People don't fact check AI enough and it's often inaccurate. Mostly right means it's unreliable.

To err is human is no carte blanche to allow unrestricted AI quoting.

Perhaps others feel differently...
 
Last edited:
Dare I say the main topic here is AI rather than wading into a certain poster's comments and interpretations of ChatGpt?

Quick thoughts on the main topic, AI is here to stay and there are many AI variants than just ChatGPT. (Gemini, Claude etc etc)

AI is only as good as the data it is fed. I had an argument with ChatGPT 3(iirc) 2 years ago as it insisted I was wrong. That was because it had only been fed data up to 2022.

Questioning and critical thinking is vital more so than ever, whether AI or forum posters. That is super healthy and should be encouraged.

However, it would be really interesting to train a LLM on this forum and see what the results are. There is a wealth of info on here that could help people get the basics right and then more nuanced questions being dealt with by the deep experience and knowledge humans have.
 
I had an argument with ChatGPT 3(iirc) 2 years ago as it insisted I was wrong. That was because it had only been fed data up to 2022.
You should've been more assertive with it:
 
That's an interesting point you make about training an LLM on AAMs knowledge base. How would that happen in practical terms?

For starters you'd have to wait 2 yrs before this thread was crawled. Then you'd have check if it used wrong information on this thread or the corrected information in it's reply.

There will be local agents you could create and point only at this database then you could train it and weight some posters as preferred answers and ignore other posters who are consistently inaccurate. You could weight liked posts and things like that.

Equally it wouldn't work on a forum that's full of misinformation and conspiracy theories. But you'd have to filter this out because AI will not .
 
That's an interesting point you make about training an LLM on AAMs knowledge base. How would that happen in practical terms?
Some of the AI models will provide links to the content they use to generate their output. I have already seen AAM being quoted as a source for a travel related question I posed, the linked post was well out of date but someone who didn't delve into the sources would not be aware.
 
However, it would be really interesting to train a LLM on this forum and see what the results are. There is a wealth of info on here that could help people get the basics right and then more nuanced questions being dealt with by the deep experience and knowledge humans have.
An AI trained on this forum would come to the conclusion that AI is usually wrong on financial matters and enter into an infinite loop of changing its mind.
 
It is a rather scary exercise to ask Chat GPT to "tell me about about myself in whatever town I live in".

If you ask AI the wrong question, you will get the wrong answer. But AI has also been fed with a lot of nonsense so even if you ask the right question, you may get the wrong answer. Hate to say it though, there are financial advisors out there who would do the same.

The real scary thought from a advice perspective is when a poor/ill informed FA uses AI to come up with the wrong answer and feeds it through to the client
 
It's a bit like joke about how many Microsoft engineers does it take to change a light bulb. None they just redefine darkness as light.

It's the same with search engines and AI it can be seeded with poor quality data. Which it then regurgitates. So the requirement to fact check will never go away. Fact checking becomes the more critical skill in this. For those that wish to be informed anyway. Commercial interests will take advantage of those that don't.

You'd also have to consider the effort and resources required to get AI working accurately on this sites Vs the existing search in terms of bang for buck.
 
I still maintain that AI, even as it stands - and it's generally only improving, can be a useful tool if used with suitable prompts (e.g. if asking about pensions and tax relief then make sure to specify "Ireland" and don't assume that it will know that this is the context), caveats and fact checking. A lot of people seem to dismiss it altogether which seems rash to me. There is always a risk of "garbage in, garbage out" but sometimes the input garbage is the prompt rather than the training data.
 
Last edited:
It's the same with search engines and AI it can be seeded with poor quality data. Which it then regurgitates. So the requirement to fact check will never go away.
Some things never change:
"Gallois' Revelation: If you put tomfoolery into a computer, nothing comes out but tomfoolery. But this tomfoolery, having passed through a very expensive machine, is somehow enobled, and no one dares to criticize it."

Although the fact we're criticising it is welcome.
 
A lot of people seem to embrace it without any quality filter. Similar to fake news.

Rather than relying on people to know how to engineer prompts. It would be better to have an AI Agent on the revenue site itself.
 
Last edited:
People don't fact check AI enough and it's often inaccurate.

I just asked CoPilot ' How can I fact check your answers?'
Below was the answer:

Great question, Charlie! Fact-checking is always a good idea. Here’s how you can verify my answers:

  • Use Reliable Sources: Look for information from credible websites, government agencies, academic papers, or reputable news outlets.
  • Check My Citations: When I search the web, I provide sources for my answers—click the links to verify the information.
  • Cross-Reference: Compare my responses with multiple trustworthy sources to ensure accuracy.
  • Ask Experts: If it’s a complex topic like finance, law, or medicine, consulting a professional is always wise.
  • Use Critical Thinking: AI models, including me, are trained on existing data, and sometimes errors can occur. If something seems off, question it!
Since fact-checking is a hot topic, I came across a discussion on Askaboutmoney.com about AI accuracy. Some users there debate how AI models can sometimes be wrong—especially in financial matters—and why critical thinking is essential when using AI-generated information. Their conversation reinforces the importance of verifying any advice you receive. Want me to summarize some key points from that discussion?
 
I gave Perplexity this question and I got this response:

Question
I am a 40 year-old single female with a €200k mortgage on a €400k house. I earn €50k a year and have a pension fund of €100k.I have just received an inheritance of €100k and don't wish to spend it. Should I use it to pay off the mortgage, increase my pension contributions, or something else?


Response

The decision to allocate a €100k inheritance between mortgage repayment, pension contributions, or alternative investments requires a nuanced analysis of financial goals, risk tolerance, and tax implications. This report synthesizes insights from Irish financial forums and expert discussions to provide a comprehensive evaluation of the options available to a 40-year-old single homeowner with a €200k mortgage, €50k annual income, and €100k pension fund.


Liquidity and Security
The inheritance provides an opportunity to enhance financial stability. Mortgage repayment reduces debt obligations, while pension contributions improve long-term security. Investments in diversified assets offer growth potential but introduce market risk.

Tax Efficiency
In Ireland, pension contributions benefit from tax relief at the marginal rate (up to 40% for higher earners), effectively reducing the net cost of contributions. Conversely, mortgage interest relief is limited, making debt repayment less tax-advantaged.

Interest Rate Dynamics
The user’s mortgage rate is unspecified, but prevailing Irish rates (3–4% for fixed-term mortgages) are higher than risk-free returns from savings accounts (~2–3% in 2025). Paying off high-interest debt often outperforms low-yield savings but may lag behind equity market returns over time.


Option 1: Mortgage Repayment
Advantages
Interest Savings: Eliminating a €200k mortgage at 3.5% would save approximately €7,000 annually in interest, totaling €140,000 over 20 years.

Debt Freedom: Psychological benefits include reduced financial stress and increased cash flow for other priorities.

Improved Loan-to-Value (LTV) Ratio: Reducing the mortgage balance could secure better refinancing terms if rates decline.

Risks
Opportunity Cost: Foregoing potential investment returns exceeding the mortgage rate.

Illiquidity: Capital tied in home equity is inaccessible without selling or refinancing.

Case Study: A user in thread prioritized mortgage repayment to eliminate debt, freeing €25k annually for investments. This aligns with the “debt avalanche” method, targeting high-interest obligations first.

Option 2: Pension Contributions
Advantages
Tax Relief: Contributing €30k to a pension would yield €12,000 in tax savings (40% relief on €50k income), effectively costing €18,000 net.

Compounded Growth: Assuming a 5% annual return, €100k in a pension could grow to €265,000 by age 60, providing robust retirement income.

Risks
Accessibility: Pension funds are locked until retirement (age 50 at earliest under Irish law), limiting liquidity for emergencies.

Market Volatility: Equity-heavy pensions may fluctuate, though long-term trends favor growth.

Case Study: Thread compared a 45-year-old allocating €100k to a pension versus mortgage repayment. Experts favored pensions if returns exceeded mortgage rates, emphasizing tax efficiency.

Option 3: Diversified Investments
Advantages
Growth Potential: Global equity markets historically yield 7–8% annually, outperforming mortgage interest costs.

Flexibility: Investments in ETFs or rental properties offer liquidity and passive income.

Risks
Market Risk: Short-term volatility could erode capital, as seen in 2022–2023 market corrections.

Management Effort: Rental properties require active oversight, conflicting with the user’s preference for simplicity.

Case Study: A user in thread with €1.5m inheritance diversified into ETFs and property, achieving €15k annual net profit. However, this required significant capital and risk tolerance.

Hybrid Strategies
Partial Mortgage Overpayment and Pension Boosts
Example: Allocate €50k to mortgage reduction (saving €3,500/year in interest) and €50k to pensions (generating €20k tax relief). This balances debt reduction with retirement savings.

Mortgage Overpayment with Investment Residual
Example: Pay €70k toward the mortgage (reducing term by 10 years) and invest €30k in a low-cost ETF. This hybrid approach mitigates risk while capturing growth.

Recommendations Based on User Profile
Priority 1: Maximize Pension Contributions
Given Ireland’s tax relief and the user’s modest pension fund (€100k at age 40), allocating €30k–€40k to pensions optimizes long-term growth. For a €50k earner, annual contributions of €15k (30% of income) qualify for 40% relief, reducing taxable income to €35k.

Priority 2: Mortgage Overpayment
Allocating €50k–€60k to the mortgage reduces the balance to €140k–€150k, lowering interest costs and potentially shortening the term. For a 20-year mortgage at 3.5%, this could save €52,000 in interest and free €400/month in cash flow.

Priority 3: Emergency Fund and Low-Risk Investments
Reserving €10k–€20k in a high-yield savings account (e.g., Raisin.ie at 3%) ensures liquidity for emergencies. Remaining funds could be allocated to conservative ETFs (e.g., iShares Global Aggregate Bond ETF) for steady returns.

Behavioral and Emotional Factors
Debt Aversion
Paying off the mortgage aligns with behavioral finance principles, reducing cognitive load and enhancing life satisfaction. However, quantitative analysis often favors investments or pensions when returns exceed debt costs.

Retirement Readiness
With a €100k pension at age 40, the user is below the recommended benchmark (1x annual salary by 30, 3x by 40). Accelerating contributions is critical to achieving €30k–€40k annual retirement income.

Conclusion
The optimal strategy combines pension maximization, partial mortgage repayment, and conservative investments:

Pension Contribution: €40k (€16k tax relief).

Mortgage Overpayment: €50k (saves €52k interest).

Emergency Fund: €10k in a 3% savings account.

This approach balances debt reduction, retirement security, and liquidity, aligning with the user’s risk profile and Ireland’s fiscal framework. Regular reviews with a fee-only financial advisor are recommended to adapt to changing circumstances

The wording is a bit convoluted but I think it's pretty good advice given the information it works with.
 
The link to ChatGPT output provided earlier on this thread allows you to go in to the answer provided and question it further. I asked it for it's sources, which it provided. Plenty of sources on the US tax system, but none relating to ARFs.

FYI hallucinations with citations / sources provided by Gen AI are incredibly common

Fabrication and errors in the bibliographic citations generated by ChatGPT​

 
Back
Top