Chat GPT，实用工具而已，不必盲目崇拜

人们害怕自主、自决式人工智能会导致失控。Chat GPT、Ernie和Xiaoice现在发展到了什么程度？对未来的财富管理意味着什么？

推荐使用GPT中文版，国内可直接访问：https://ai.gpt86.top 

你还是事事亲力亲为，或是已在使用聊天机器人？人工智能助手Chat GPT3使撰稿人面临数字竞争，GPT4也将图形艺术家置于同样的竞争处境。我们现在是否应该担心GPT5会进行网站编程，GPT6会经营公司，GPT7会统治世界？

我们对基于规则的聊天机器人已经习以为常，许多企业网站上都使用了这类聊天机器人。然而，下一代（大型语言模型）的进展让专家和外行都感到惊讶，甚至被炒作称为“杀手级应用”。

串联单词

人工智能的发展速度确实惊人。技术史上有革命性的发明诞生时，总有人担心自己会失去工作、健康、自主决定权、权力和存在的理由。在向聊天机器人时代过渡的过程中，情况如出一辙。

乍看之下，人工智能生成文本的速度惊人，且内容易于理解——既非原创，也非创造性结果，基本上只是根据概率原理将单词串联起来，简而言之就是回声机的结果。

文本中没有谷歌或百度等搜索引擎的来源引用，因为文本由数百万个数据片段组成。有时，这会导致出现“幻觉”、人为制造的谎言、听起来可信的反社会假新闻。

这对于快速、廉价的内容创作可能非常有用，但在财富管理和金融理财规划方面却可能是巨大的危险：具有代表性的各种行动建议并不能形成经过深思熟虑后解决客户具体问题的个体概念。

机器人尚不能与人类竞争

有一点很明确：范围越小，结果越好。因此，对于结构化、网络化问题解决技能，如金融理财规划方面通常需要的技能，机器人还不是人类的竞争对手。情感、沟通和同理心仍是人类独有的。但专家对比了聊天机器人技术的颠覆性潜力与大约30年前引入商业互联网的情况。

在人工智能的使用方面，我们仍处于“试错”阶段。一旦有关数据保护的某些道德和法律问题得到澄清，我们也已学会将聊天机器人作为个人助理或“智能伙伴”，届时各行各业的工作方式将随之发生巨变。虽然我们今天认为使用互联网是一件很自然的事情，但这项技术历经多年才得到人们的认可并正常运行。人工智能亦然如此。赫拉克利特早在2500多年前曾有言，“宇宙中唯一不变的就是变化”……

使用聊天机器人作为陪练

例如，MS Office Suite中实现了这项技术，微软的最新版本称为“Microsoft 365 Copilot”——聊天机器人始终作为导航器或副驾驶出现。所有重复性工作皆可自动化，潜在应用无限。用户必须学会如何使用聊天机器人，如根据任务、目标和标准向聊天机器人指定请求，即所谓的“提示”，确定格式，然后进一步通过反馈和批评进行微调。标准是能够自动回答简单问题。 

但问题是，聊天机器人是否也会进入下一阶段：谁对投资建议负责？聊天机器人能否成功为可持续的客户关系建立信任？ 可以想象，聊天机器人首先可以简化顾问的准备工作，从而有助于进行财富管理。人类现在是且仍然是控制者和决策者，聊天机器人只是工具和陪练。因此，金融理财规划软件系统迟早也会实现人工智能，只是时间问题而已。

本文是纯人类脑力劳动的成果。Chat GPT Spring表示。

People fear the loss of control through autonomous, self-determined artificial intelligence. Where does the development of Chat GPT, Ernie and Xiaoice currently stand and what does it mean for the future of wealth management?

Are you still working by yourself or are you already using chat bots? With the AI assistant Chat GPT3 copywriters have got digital competition, and so have graphic artists with GPT4. Should we now fear that GPT5 will program websites, GPT6 will run a company, and GPT7 will take over the world?

We have become accustomed to rule-based chatbots, they are used on many corporate websites. However, the progress of the next generation (LLM) has surprised experts and laymen alike and is even being hyped as a "killer application".

Stringing words together

The speed of development of artificial intelligence is indeed stupendous. Revolutionary inventions in the history of technology have always been accompanied by fears that people would lose their jobs, their health, their self-determination, their power and their raison d'être. The transition to the chatbot age is no different.

As amazingly quick and comprehensible as the artificially generated texts appear at first glance - they are neither original nor creative, but basically a stringing together of words according to the principle of probability, in short: the result of an echo machine.

There can be no source citations as with the search engines like Google or Baidu, since the texts are composed of millions of data snippets. This sometimes leads to "hallucinations," artificially generated untruths, plausible-sounding, sociopathic fake news.

What might be very useful for fast and cheap content creation, might be a big danger in wealth management and financial planning: A cross-section of different recommendations for action does not make a well-thought-out individual concept that addresses a customer's specific problems.

Robots are not yet a competition for humans

One thing is clear: the narrower the scope, the better the results. For structured, networked problem-solving skills, such as those typically required in financial planning, robots are therefore not a competition yet. And where emotions, communication and empathy count, humans remain unique. But experts compare the disruptive potential of chatbot technology with the introduction of the commercial internet some thirty years ago.

When it comes to the use of AI, we are still in the "trial and error" stage. As soon as certain ethical and legal questions about data protection have been clarified and we have learned to use chatbots as personal assistants or "smart companions”, the way we work in our industry will also undergo a massive transformation. As natural as the use of the internet seems to us today, it took several years for the technology to be accepted and to function properly. It will be no different with AI. "The only constant in the universe is change," said Heraclitus more than 2500 years ago…

Using chatbots as sparring partners

The implementation of this technology in the MS Office Suite, for example, makes Microsoft call the latest version "Microsoft 365 Copilot" - the chatbot is always there as a navigator or co-pilot. All repetitive work can be automated, and the potential applications are endless. Users must learn how to make use of chatbots, for example by specifying requests to the chatbot, so-called "prompts" according to tasks, goals, and criteria, determining the format and then, in further steps, doing the fine-tuning through feedback and criticism. Automated responses to simple questions will become the standard.

The question will be whether chatbots will also make it to the next level: Who is liable for investment recommendations? Will they succeed in building trust for a sustainable customer relationship?  It is conceivable that they will first be used as an aid in wealth management by simplifying the preparations for advisors. The human being is and remains the controller and mastermind, the chatbot is only a tool and sparring partner. It is therefore only a matter of time before AI is also implemented in software systems for financial planning.

This text was elaborated with purely human brain work. Says Chat GPT Spring.