AI: The Financial Adviser Is Next

I once heard that a financial adviser is a man who takes your money and then invests it for you until it is all gone. This may be unfair. Plenty of financial advisers are honest, competent people who genuinely help their clients. Even so, the joke survives because it captures something important. Financial advice is expensive. Whether the cost is visible or hidden, somebody is always taking a cut. Advisers charge fees. Fund managers charge fees. Brokers charge fees. Entire armies of intermediaries stand between ordinary people and their own money. For generations we were told this was necessary. Investing was complicated. The stock market was mysterious. Ordinary people lacked the expertise needed to analyse companies, read accounts, understand macroeconomics, or evaluate risk. The experts would do it for us. Artificial intelligence is rapidly destroying this argument.

Though still under age, I have been using one of my grandmothers as a front in order to invest since 2023. The arrangement may not satisfy every compliance officer in the City of London, but it has been educational. My gold and silver exchange-traded funds have not performed quite as spectacularly as I expected. Even so, my overall gains have been somewhere around fifty per cent. I did not obtain these results by hiring a wealth manager. I did not spend years studying for a finance qualification. I did not join a hedge fund. Instead, I spent many of the late night hours I might have given to Pornhub talking to artificial intelligence.

The machines explained concepts I did not understand. They taught me how to read company reports. They helped me compare businesses. They suggested industries worth investigating. They challenged my assumptions. Most importantly, they enabled me to perform research at a speed that would have been impossible only a few years ago.

This does not mean artificial intelligence is a magic money machine. Anyone who believes that is asking to become poorer. The Daily Telegraph recently reported on experiments in which several leading AI models were allowed to manage hypothetical investment portfolios. The results were mixed. Some systems performed terribly. Anthropic’s Claude Sonnet reportedly lost nearly sixty per cent of its capital. Google’s Gemini also performed badly. ChatGPT made a modest profit, while Elon Musk’s Grok roughly broke even.

At first sight, these results seem disappointing. If artificial intelligence is supposed to be revolutionary, why is it not producing Warren Buffett-like returns? The answer is that this misses the point. Artificial intelligence does not need to become the greatest investor in history to transform investing. It merely needs to become cheaper and almost as competent as the average professional adviser. That threshold is already being crossed.

One reason is scale. A human analyst can read perhaps a few reports each day. An artificial intelligence system can process thousands of pages in minutes. It can compare years of financial statements. It can examine earnings calls. It can identify competitors. It can search regulatory filings. It can cross-reference information from dozens of sources simultaneously.

The Telegraph article highlighted one striking example. ChatGPT identified Credo Technology Group as a potential beneficiary of growing demand for internet infrastructure. The company was obscure enough that even experienced investors had largely ignored it. Since then, the stock has risen dramatically. This does not prove that AI possesses supernatural forecasting powers. It proves something more interesting. Machines are becoming increasingly good at finding needles in haystacks.

The financial industry has traditionally depended on information scarcity. The professional’s advantage came from having access to data that ordinary people lacked, or having enough time to analyse it properly. Both advantages are disappearing. Public information is now abundant. Artificial intelligence can digest it faster than any human being. The result is that much of the traditional value offered by advisers is becoming difficult to justify.

This process is not confined to amateur investors. The most revealing part of the Telegraph article was not that ordinary people are experimenting with ChatGPT. It was that some of the world’s largest hedge funds are already using similar systems. Man Group, one of the largest hedge funds on earth, has reportedly allowed language models to propose investment ideas that have eventually been approved and implemented. Balyasny Asset Management has put AI to work analysing vast quantities of corporate information. Bridgewater Associates and Two Sigma have invested heavily in machine learning systems.

This matters because hedge funds are not charities. They do not adopt technologies because they are fashionable. They adopt technologies because they believe those technologies provide an advantage. When billions of pounds are at stake, sentimentality tends to disappear.

The emerging pattern is obvious. Artificial intelligence is not replacing investors entirely. It is acting as an intellectual force multiplier. A fund manager who previously tested two or three ideas each day can now examine hundreds. Research that once took days can be completed in minutes. Entire categories of repetitive analytical work can be automated.

For now, humans remain in control. The important phrase is “for now.” Every profession confronted by automation tells itself the same comforting story. The machines will assist us. The machines will complement us. The machines will make us more productive. This is often true during the early stages. Then the machines improve. The history of technology is filled with occupations that believed themselves indispensable until they suddenly weren’t. Telephone operators disappeared. Typists disappeared. Travel agents mostly disappeared. Countless clerical roles vanished. The financial adviser is beginning to look vulnerable for the same reason.

The traditional adviser performs three main functions. He gathers information. He analyses information. He communicates conclusions. Artificial intelligence already performs all three tasks competently. Gathering information is trivial. Analysing information improves every month. Communicating conclusions is literally what large language models were designed to do. The remaining human advantages are trust, emotional reassurance, and legal responsibility. Those are not trivial advantages. Many people want someone to blame when investments go wrong. Many people need emotional support during market crashes. Many wealthy clients prefer dealing with another human being. But these are shrinking islands in an expanding technological sea.

The younger generation increasingly turns to artificial intelligence first whenever they need information. They ask ChatGPT about careers. They ask ChatGPT about relationships. They ask ChatGPT about programming. They ask ChatGPT about health. It would be strange if investing remained exempt.

Indeed, I suspect the financial advice industry faces the same long-term problem confronting much of the professional middle class. The product they sell is knowledge. Artificial intelligence is reducing the scarcity value of knowledge. This does not mean everyone will become rich. Burton Malkiel famously joked that monkeys throwing darts at newspaper stock tables could match professional fund managers. The efficient-market hypothesis remains a serious challenge to anyone claiming to possess a permanent investing edge. Artificial intelligence may ultimately prove no better than humans at predicting the future. Even if that turns out to be true, the consequences remain enormous.

If AI can equal the average adviser while charging almost nothing, why would anyone pay thousands of pounds for traditional advice? The question answers itself. Since the end of the last Ice Age and our transition from hunter-gatherer to settled life, economic progress has involved replacing expensive human labour with cheaper tools. Artificial intelligence is simply extending that process into domains that educated professionals once assumed were protected. The taxi driver who dispenses stock tips in the pub may survive for a while longer. The financial adviser probably will too. But when a teenager with a laptop can access analytical capabilities that would have astonished Wall Street twenty years ago, it becomes difficult to believe that the existing order will survive unchanged. The machines may not always know which stock will rise next. What they increasingly know is how to perform much of the work for which financial advisers currently charge a premium. That fact alone should make every adviser in Britain nervous.