Why ChatGPT could be a step backwards?

The launch and clever marketing of OpenAI’s ( OpenAI ) ChatGPT (#chatgpt ) has led to heated debate with opinions ranging from it being the greatest invention ever to AI failing to answers simple questions and being too “robotic”. What has been missing in the discussion so far is the risk of possibility that ChatGPT could lead a large number of people into believing that our world is deterministic. That there is only one correct answer.

Humans have always sought information or advice. In the earlier days it was religious priests telling them “what is the best way to live their lives” which became open to personal-interpretation once the religious texts became readily available. As printing became cheaper, people were able to select not only the kind of information they were looking for by selecting books that dealt with the specific topics, but also the answers that suited their thinking by following certain authors. In recent times, people have adapted more or less similar/ analogous version through different mediums (like podcasts, MasterClass , TED Conferences , social media influencers on platforms like Instagram )

Google ( 谷歌 ) changed that dynamics by ranking the results based on its understanding of which answers are most relevant to the person’s query. But it still shows all the results, and the users are able to see that there are different answers to the same questions. For its part Google did a good job in distinguishing between fact and opinions.

ChatGPT is an innovation that offers conversational search functionality. The user asks a questions (e.g. what to do when one discovers a mole on their skin?) and the ChatGPT produces response. See screenshot for ChatGPT's response to that particular question.

The way ChatGPT is being marketed, there is a risk that an average user will get the impression that this is the "final wisdom" based on thousands of documents available on the internet. Which is not going to be true in a majority of the cases.

A quick scan of Google search informs even an average user that there are different points of views. The user can then check the website addresses, their credibility and evaluate which other sites have similar or opposing views.

ChatGPT is trivializing the complex and uncertain nature of many of the issues we face. Taking a deterministic view to important things like health, business decisions or policy making is not only simplistic but could be counter-productive. One of the cornerstone of today's world is people having opposing views and often being correct in their own ways. When strategic consultants like 麦肯锡 , 波士顿谘询公司 , 贝恩公司 provide different advice to similar problem, how can ChatGPT claim to have modelled the information so perfectly that there is one answer? Would the next version of ChatGPT make the gathering of world leaders in World Economic Forum #davos useless?

The real risk of ChatGPT is not that it will become too powerful and control our world. A more realistic scenario is that in its quest to oversell its potential, we will end up with a population that is less informed and wrongly advised than it would be without ChatGPT.

Hopefully, the discussion would move away from the ChatGPT's power to the "modelability" of information, knowledge and actionable wisdom. We live in a probabilistic world. It will be wise to realize the marketing gimmick from the reality.