想用人工智能管理个人财务？提问可得讲究技巧。
内容总结：
人工智能（AI）正逐渐渗透个人理财领域，但专家提醒投资者需谨慎使用。尽管AI能够快速解释投资术语、提供基础策略建议，甚至筛选市场数据，但由于当前大语言模型存在信息准确性不足、假设不一致等问题，盲目依赖AI可能导致投资决策失误。
Betterment财务规划总监尼克·霍尔曼指出，许多投资者缺乏提出精准问题或使用专业术语的能力，而AI生成的答案时常出现错误或自相矛盾。OpenAI也在服务条款中明确提示，其输出内容“未必始终准确”，不应作为唯一事实来源或专业建议的替代。
尽管存在风险，AI在理财中的应用仍在快速普及。Experian调查显示，近半数受访者正尝试或考虑使用生成式AI管理财务，Z世代和千禧一代的使用率尤其高。目前，AI更适合辅助投资者理解概念、生成初步投资思路，而非直接管理资金。
纽约大学斯特恩商学院教授瓦桑特·达尔强调，投资者需具备独立验证AI建议的能力。如年轻投资者约翰·尼尼亚通过结合自身金融分析经验与AI的数据整合能力，成功优化了投资决策。未来，随着技术发展，AI或将为普通投资者提供更可靠支持，但目前仍需通过券商专业服务或个人深入研究来验证AI提供的信息。
（本文基于《巴伦周刊》相关报道综合整理）
中文翻译：
想利用人工智能来管理个人财务吗？提问时需字斟句酌。
除非明确自己的需求，否则人工智能可能会给出错误或具有误导性的投资策略建议。
在个人理财领域，人工智能是一把双刃剑，最适合那些目标明确的资深投资者使用。机器人投顾Betterment的财务规划总监尼克·霍尔曼指出，人工智能虽能快速生成答案，但许多投资者根本不知道应该提出哪些问题或使用哪些专业术语。即便如此，聊天机器人有时仍会给出错误或自相矛盾的答案。
他表示："人工智能无疑具有重塑金融服务的潜力，但目前尚不具备独立管理资金的能力。"他补充说，当前的大语言模型"表现令人惊叹，但对于管理养老钱这般重要的决策，仍需保持谨慎。"
他观察到的另一个问题是假设前提不一致。例如，大语言模型可能在一次计算中采用2%的通胀率，下一次却使用3%；或在某次大学费用预测中包含书本费，下次却只计算学费。当联系OpenAI（ChatGPT的母公司）置评时，其通讯团队援引服务条款回应称"输出内容未必始终准确"，用户不应将输出结果"视为唯一真相来源或专业建议的替代品"。
美国公众似乎并未重视这些警告。皮尤研究中心6月发布的调查显示，自2023年夏季以来使用该工具的美国成年人比例增长约一倍，34%的受访者表示至少使用过一次。益博睿去年发布的调查发现，在2011名受访者中，近半数已使用或正考虑使用生成式AI工具管理财务；Z世代和千禧世代中这一比例分别达67%和62%。
诚然，ChatGPT在解释投资术语或基础策略时颇具价值，例如解读60/40投资组合的优势或美元成本平均法原理。它能快速列出按资产管理规模排序的标普500指数ETF龙头，或罗列市场主要科技股名单。但即便如此，用户仍需通过可靠渠道复核信息——因ChatGPT仍可能提供错误内容。托数字新闻中心今年初测试8款生成式搜索工具（含ChatGPT）的研究显示，超过60%的提问获得了错误答案。该研究还发现，聊天机器人倾向于对无法回答的问题给出错误或猜测性答案，而非拒绝回答。
纽约大学斯特恩商学院数据科学教授瓦桑特·达尔指出，要真正从工具中获取价值，"必须具备评估ChatGPT输出内容的能力"，尤其当用户打算根据其建议采取实际行动时。
现年25岁、来自纽约长岛的对冲基金总监约翰·尼尼亚正是具备这种能力的投资者。他16岁起参与市场投资，大学毕业后曾在Mobius Investments担任金融分析师。他日常的工作包括筛选关键指标股、分析企业核心竞争力及竞争对手。近年来他开始借助ChatGPT辅助工作。
现居迪拜的尼尼亚表示："它能用更易理解的方式解释概念并接受追问"，该AI工具还能快速调取市盈率、潜在投资回报、企业高管信息等数据。"获得这些远超人工效率的信息后，你可以对比三四家公司并做出更优决策。"
尼尼亚称该工具实用到足以支撑其个人投资。此前特朗普总统与美联储主席鲍威尔关系紧张时期，他曾用ChatGPT查询美国三大住宅建筑商名称及财务数据，最终根据估值与营收比率等财务指标，投资了估值最低的一批股票。
终有一日，AI工具或能让我们都像尼尼亚一样投资——目前已有许多公司试图通过该技术为普通投资者赋能。例如投资软件Public.com推出AI功能Alpha，可通过新闻提醒提供资产摘要、组合资产咨询聊天服务等。其首席运营官斯蒂芬·赛克斯表示，该工具常用于解读财报电话会议和分析师群体情绪。
"它专注于合成散户投资者最想了解的企业核心信息。"赛克斯说道。
因此，虽然AI终将能协助投资者做出关键投资决策，但目前最大价值在于帮助理解概念和获取初步投资思路。所有这些都需要通过受信任的券商或自主研究进行验证。
欢迎来信反馈至editors@barrons.com
英文来源：
Want to use AI for your personal finances? Choose your questions carefully.
Unless you know what you are looking for, artificial intelligence can give wrong or misleading on investment strategies.
When it comes to personal finance, artificial intelligence is a two-edged sword best wielded by experienced investors who know exactly what they want.
It has the ability to get you quick answers, but many investors don’t know which questions to ask or the specific terminology to use in the first place, says Nick Holeman, director of financial planning at robo-advisor Betterment. And even then chatbots will sometimes give wrong or inconsistent answers.
“There’s no question that AI has the potential to reshape financial services, but it’s not ready to manage your money on its own just yet," he says. He adds that today’s large language models (LLMs) are “impressive, but with decisions as important as your financial nest egg, you should be careful."
Another problem he has seen is inconsistent assumptions. For example, the LLM might use 2% inflation in one calculation, and in the next use 3% inflation, or it might include the cost of books in one college forecast but only tuition in the next. When contacted for comment, the communications team at OpenAI—the parent company of ChatGPT—pointed to the company’s terms of use, which includes that “output may not always be accurate" and that users shouldn’t rely on the output “as a sole source of truth or factual information, or as a substitute for professional advice."
The American public seems unlikely to heed such cautions. The share of U.S. adults who use the tool has roughly doubled since the summer of 2023, with 34% saying that they have used it at least once, according to a survey published by the Pew Research Center in June. A survey from Experian published last year found that nearly half of its 2,011 respondents have used or are considering using generative AI tools to help with managing their finances; 67% of Gen Zers and 62% of millennials say they already do so.
To be sure, ChatGPT can come in handy when it comes to explaining investing terms or basic strategies, like the value of a 60/40 portfolio or how to dollar-cost average. It can spit out the top S&P 500 exchange-traded funds (ETFs) by assets under management or give you a list of some of the leading technology stocks on the market. But even in this case, you should double-check your work with a reliable source, since ChatGPT can still share incorrect information. In a study published earlier this year, the Tow Center for Digital Journalism tested eight generative search tools including ChatGPT and found that the bots offered up incorrect answers to more than 60% of queries. The study also found that the chatbots tended to be bad at declining to answer questions they didn’t have the correct answers to, and instead offered wrong or speculative answers.
Vasant Dhar, a professor of data science at the NYU Stern School of Business, says that to really get value out of the tool “you have to have the chops to evaluate what ChatGPT is telling you," especially if you’re actually going to act on any of its recommendations.
Some investors, like 25-year-old John Ninia, originally from Long Island, N.Y., do have those chops. Ninia has been putting money to work in the market since he was 16, and when he graduated college, he worked as a financial analyst at Mobius Investments until more recently becoming a director at the hedge fund. His days consist of scanning for stocks with key metrics, understanding what makes a company stand out from the crowd and who its competitors are. In recent years, he’s tapped ChatGPT to help.
“It can explain it to you in a way that you can more easily understand and you can ask questions," says Ninia, who is based in Dubai, adding that the AI tool can also easily pull up price-to-earnings ratios, potential returns on investments, information about company executives and more. “Once you have this information that it can gather much quicker than anyone could, you can compare the three or four companies with each other and make better decisions."
Ninia says the tool is so useful, he has used it for personal investments. During the recent tension between President Donald Trump and Federal Reserve Chair Jerome Powell, Ninia wondered whether the president’s pressuring the Fed to lower rates would work and—if so—whether home-building and buying will rev up as people look for lower rates. He asked ChatGPT for the names of the three largest home builders in the U.S., along with their financial information. He ended up investing in a group of those stocks that had the lowest valuations compared with the revenues they were generating, among other positive financial measures.
Eventually, AI tools may be able to help us all invest like Ninia—and many companies are already looking to tap the technology to give everyday investors a leg up. Investing app Public.com, for instance, has Alpha, which provides investors with AI-powered summaries and recaps of assets via news alerts, a chat function to ask questions about assets in your portfolio and more. Stephen Sikes, Public.com’s chief operating officer, says common uses of the tool include recapping earnings calls and understanding broader sentiments from across the research analyst community.
“It’s very focused on synthesizing the most common pieces of information retail investors want to understand about a company," Sikes says.
So while AI may one day be able to help investors make crucial decisions about their investment portfolios, right now it can be most helpful to help you understand concepts and get initial investing ideas. All of which need to be backed, either by the brokerage companies you are trusting to manage your money or via your own research.
Write to editors@barrons.com