I’ve spent years building a solid second income stream from a spread of UK dividend stocks, but today I tried to do the same job in seconds.
I did this by asking ChatGPT to help me design a portfolio of FTSE 100 stocks that would pay me a passive income £2,000 a month, or £24,000 a year. For guidance, I suggested choosing half a dozen shares that produced an average yield of 5% a year (higher than the average blue-chip yield of 3.25%). To do that would require £480,000.
It’s a big sum, but a realistic long-term goal for anyone who invests consistently over decades in a Stocks and Shares ISA or Self-Invested Personal Pension (SIPP).
FTSE 100 dividend stock picks
I’ve played around with ChatGPT before, but would never rely on it to buy stocks. Its choices are often based on out-of-date info and understate the potential risks.
The chatbot is fast though. In seconds, it came back with insurer Legal & General Group, saying it “offers one of the strongest yields in the index at around 8%”. The trailing yield’s actually 9%, but I’ll let that pass.
It’s true that the “pension and insurance heavyweight” has deep cash generation and a long track record of paying generous income. It’s also true that its shares have gone nowhere in a decade. ChatGPT doesn’t mention that.
Its second pick was another stock in the financial sector, wealth manager M&G, another big yielder at roughly 8%. Its shares have performed better than Legal & General’s, but diversification’s key. Of the two, I’d consider M&G first.
National Grid shares worry me
The chatbot then moved on to hardy income perennial National Grid (LSE: NG), saying the energy utility “gives the portfolio stability”.
ChatGPT waxed lyrical saying: “The dividend rises a touch above inflation, the income is predictable, and the yield is high enough to make a real difference”.
It then gets the yield totally wrong, claiming it pays income of 6% a year, when it’s now just 4.1%. Big difference! Personally, I’m wary of National Grid. It has to invest around £60bn by March 2029 to upgrade and modernise the electricity grid to handle renewables.
In May 2024, the board shocked investors with a massive £6.8bn rights issue. Given how infrastructure projects often run over, I can imagine it doing that again, risking dilution. Plus it already has more than £40bn of debt. I think National Grid’s riskier than ChatGPT pretends. Not for me.
It also picks consumer giant Unilever, which I recently sold, underwhelmed by its growth prospects and 3.25% dividend yield.
ChatGPT threw oil giant Shell into the mix, says its 3.9% yield is “supported by massive free cash flow, share buybacks and a willingness to return cash to shareholders”, and finished what what it called a “wildcard” in telecoms giant Vodafone.
It said the yield‘s huge, around 10%. Simply not true. the dividend was recently slashed in half and the yield’s now 4.27%. As it admits, ChatGPT can make mistakes.
While there are some decent names here, I won’t consider National Grid, Unilever or Vodafone. Investing is a personal business, and relying on a robot to picks stocks simply doesn’t cut it. Proper retirement planning requires more than a few seconds fiddling about on ChatGPT.
