Hallucinations can turn the promise of AI into a bad trip. Here’s how to separate slop from reality

As artificial intelligence becomes increasingly embedded in investment research and decision-making, family offices are exploring how to harness its potential while managing its risks. A recent Canadian Family Offices article examines one of AI’s greatest challenges: hallucinations, where generative AI produces convincing but inaccurate information. While AI continues to evolve rapidly, the article highlights that verification and human oversight remain essential, particularly in high-stakes environments such as wealth management.

In the article, Spencer Clark, Head of Private Markets and Thematic Investments, explains that AI can be a valuable tool for improving efficiency and accelerating preliminary research, but it should never replace rigorous analysis. At Richter, AI is used to help synthesize large volumes of investment information, while every output is independently validated before informing investment decisions. Clark notes that this is especially important in private markets, where publicly available information is often limited and generative AI may produce responses that combine accurate facts with anecdotal or entirely incorrect information.

The article reinforces that the responsible adoption of AI depends on combining trusted technology with disciplined governance and experienced professional judgment. As AI capabilities continue to advance, Spencer emphasizes that the greatest value comes from using these tools to support better decision-making, not to replace the expertise and critical thinking that clients rely on.

Read the full article on the Canadian Family Offices website.