Here’s a bold statement: relying solely on AI to build your retirement portfolio could be a risky move. But is it entirely off the table? I decided to put ChatGPT to the test by asking it to craft the ultimate Self-Invested Personal Pension (SIPP) for someone like me—someone eyeing retirement within the next decade and prioritizing dividend income. The results were… intriguing, to say the least. Let me walk you through what happened and why it’s sparked some serious questions about the role of AI in investing.
First, a bit of context: my SIPP is already packed with FTSE 100 stocks, and I’m mostly content with it. But as retirement looms closer, I’ve started second-guessing whether it’s truly optimized for my needs. Enter ChatGPT, the AI assistant everyone’s talking about. I asked it to whip up a ‘perfect’ blend of dividend-paying stocks, and here’s where it gets controversial: while its suggestions weren’t terrible, they weren’t exactly groundbreaking either.
The Good (and the Flattering):
ChatGPT kicked things off by calling my request a ‘great question.’ Nice try, AI, but I’ve learned to take its compliments with a grain of salt. Its top pick? Legal & General Group (LSE: LGEN), which it praised as a ‘high-yield financial/insurance business and one of the FTSE 100’s top dividend payers. Fair enough—I already own it, and its trailing yield of 9% is hard to ignore. ChatGPT called it a ‘compelling pick’ for stable cash generation, which aligns with my retirement income goals. So far, so good.
But here’s where it gets questionable:
When I asked for risks, ChatGPT served up three generic warnings: high yields can be tough to sustain, insurance businesses face long-term liabilities and market risks, and dividend stability depends on cash flow. True, but hardly insightful. It also failed to mention Legal & General’s lackluster share price performance—up just 8.4% in the last year and barely higher than a decade ago. Growth? Not so much. But for income, it’s still worth considering.
The Overconcentration Red Flag:
Next, ChatGPT recommended M&G and Phoenix Group Holdings, both from the financials sector and boasting super-high yields. But wait—three out of five stocks in the same sector? That’s a recipe for an unbalanced portfolio. And this is the part most people miss: AI doesn’t think critically about diversification. It’s just regurgitating data it’s found online.
The Wild Card:
Then came British American Tobacco, a solid pick with a stellar track record of income and growth. But the real head-scratcher was Foresight Environmental Infrastructure Trust, a FTSE 250 investment trust with an 11% yield. Sounds great, right? Until you realize its shares are down 15% in the last year and a whopping 48% over three. Risky? Absolutely. Yet ChatGPT didn’t even mention this.
The Expansion (and the Reality Check):
Five stocks weren’t enough, so I asked for more. ChatGPT added Shell, Unilever, National Grid, AstraZeneca, and Rio Tinto—a mix of income and growth. Not bad, but I’m not sold on Unilever and National Grid. Here’s the kicker: ChatGPT doesn’t generate its own ideas; it just scours the internet for human-researched information. And that’s the crux of the issue.
The Bottom Line:
While this portfolio isn’t terrible, it’s far from perfect. There are more exciting FTSE 100 income stocks out there, and a few of these picks I’d avoid entirely. AI can be a helpful tool, but it’s no substitute for human judgment. Investors, beware: always do your own research. Now, here’s a thought-provoking question for you: Can AI ever truly replace human intuition in investing, or will it always fall short? Let’s debate in the comments!
