In my opinion, ChatGPT should be used as an AI-assistant technology that helps people to accelerate the reserach and development processes. These processes could be related to diverse domain-specific applications such as writing software code in different laguages, writing blogs, writing songs etc.
Based on my understanding, ChatGPT is an AI-assistant that runs the google search in the background, selects the relevant knowledge, refines this knowledge, and reproduces it for people in a matter of tens of seconds. But due to the data-driven nature of AI methods, this knowledge can be biased, accurate or have errors. As a user of this technology, it is our responsibility to check and ensure whether the produced AI-assisted output is valid within the given context. If not, then we shall fix it manually.
Moreover, in terms of AI assisted design space. I believe, there is a consensus among the AI community. AI must be adopted across various domains where a human is in the loop. This could be in the form of rule-based policies etc.
ChatGPT is a prototype dialogue-based AI chatbot capable of understanding natural human language and generating impressively detailed human-like written text. It is the latest evolution of the GPT – or Generative Pre-Trained Transformer – family of text-generating AIs.
- Trained by AI and machine learning, the system is designed to provide information and answer questions through a conversational interface.
- Early users have described the technology as an alternative to Google because it is capable of providing descriptions, answers and solutions to complex questions including ways to write code, and solve layout problems and optimisation queries.
- Regarding the replacement of humans by the ChatGPT robot, the question remains for complex analytical tasks that the near or distant future will be able to answer.
For more details and information about this subject, I suggest you see the links on the topic.
ML models have been applied in drug discovery, but based on what we've read so far, ChatGPT is not able to make discoveries, at least for now. I'm a little confused by what you mean by "contribute to the prediction of practical experiments results", perhaps if you could define the "contribution" in broad terms, it could be easier to answer. But sure, ChatGPT could interpret the results of the experiments, that's a contribution too, but it would be interesting to actually check it on novel data to see if the interpretation is accurate or biased.