Image source: Getty Images
I’m looking to generate passive incoming retirement from a portfolio of high-yielding FTSE 100 income stocks. I have a few in mind but decided to throw the question out to ChatGPT.
I’ve learned from previous attempts that it’s unwise to use a chatbot to pick stocks, because they tend to hallucinate or deliver out-of-date information. So I was only doing this update to have a bit of fun. I asked it to pick five stocks – here’s what it came back with.
Top of the list was one of my pesonal favourites, wealth manager M&G, which has a trailing yield of 7.5%. It’s important to remember that ChatGPT doesn’t have an opinion of its own, it just lifts stuff from the web. And this pick was taken from an article published on The Motley Fool.
FTSE 100 dividend picks
It’s a very good article, naturally, but it was written on 5 August. Our Fool has written a lot about M&G since, so why pick a piece that’s more than three months old? Things change quickly in the investment world, so that was a bad start.
ChatGPT followed this by suggesting two more stocks in the financials sector – Legal & General Group and Phoenix Group Holdings. Both have their charms, with yields of 8.9% and 8.1% respectively.
But to pick three very similar stocks in a five-stock portfolio is a howler from a diversification point of view. It would leave the investor highly exposed to a downturn in this sector. Partly my fault for stipulating such a high yield though.
It’s next pick was a very different sector as it highlighted one of the most popular dividend stocks of all, British American Tobacco (LSE: BATS). I personally don’t buy tobacco shares, but ChatGPT wasn’t to know that. It’s been a brilliant income stock, hiking dividends every year this millennium. Today, the trailing yield is 5.82%.
Unfortunately, ChatGPT got that wrong, citing between 7% and 8%. The reason soon became clear. It trawled the info from an article (not the Fool this time) that appeared in June.
British American Tobacco shares
Basing stock decisions on out-of-date information just isn’t the way to work. ChatGPT’s better at throwing up generic information, for example, it says British American Tobacco has “strong cash flow and a long history of dividend payments”, which is true but not hugely enlightening.
At my prompting, it also warned of threats, sighting “structural decline in cigarette volumes, regulatory risks, and the shift to newer products may be slower than required”.
I think British American Tobacco’s well worth considering for those building a balanced income portfolio. But investors must do their own research.
Approach with caution
For the final pick, my erratic robot buddy chose retailer B&M European Value Retail. The info was lifted from an even older article, written before the shares crashed. They’re down 60% in the last year, and the stock has crashed out of the FTSE 100. ChatGPT didn’t mention that. B&M European Value Retail has a bumper yield of around 9%, but that’s no fun if capital’s being hammered.
I described using ChatGPT as a bit of fun, but there’s not much fun about losing money. I might play around with AI, but when investing real money, I want the human touch.