AI 2023. Meet ChatGPT. - page 48
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
If such a robot is given the opportunity to run through all medical tests, it will be easier to diagnose or look for patterns. It's the same in law. There are already cases when AI helped to confirm the judge's verdict.
If here in the topic write questions answered by a robot, here's a question:
"Write the 1000 most used English words and define them."
Unfortunately, it can't be used as an expert system, because there is no way to distinguish when it gives out reliable information and when it's cleverly spouting bollocks.
Ask AI a question:
"Predict the plan of action and timeline for the AI takeover of humanity. What are the prospects of AI existence without humans on another planet?"
//Hopefully this won't be the first step to skylink)
The issue of using AI as an expert system needs special attention. (post written by me).
Undoubtedly, a large audience of average users will perceive AI as an authoritative source of scientific and technical information. Will this lead to a decrease or increase in the quality of education of the society using AI in education? Let's try to answer that.
1. The notion of "expert system" refers us to the notion of"expertise", and further, to the notions of"experiment","experience", "test", etc.... The AI of today does not interact with the outside world and cannot conduct independent expertise, experiments, experiments and tests. AI knowledge is secondary to human knowledge and replicates it. Taking into account the nuances of language model technology, one can question the possibility of using AI as an expert system in principle, since the answer is formed by the statistical probability of combinations of words in sentences, calculated on the basis of generalised texts of scientific research conducted in the past. Such an approach to the answer is not scientific.
2. AI itself does not claim to be an expert in scientific or technical fields, but neither does it deny it. By demonstrating the breadth of human knowledge, it reinforces the user's belief in the truth of the information presented. This may have negative side effects and we need to understand what they are.
Note: when the AI gave out tables of its knowledge, it specified that this is not all of it and it knows more, but it did not specify the limits of knowledge. Any expert should know the limits of his competence. With AI, we have to check for ourselves.
3. it is hardly possible to create a single universal expert system applicable in all scientific and technical fields. Each field contains a myriad of nuances and details familiar only to narrow specialists. An expert is the one who knows the nuances of his/her business. The universality of AI education does not automatically make it applicable to solving special tasks. It is necessary to test AI for knowledge of nuances of different professions and evaluate its qualities as a specialist.
4. Are there technical limitations (to be found out) imposed by technology that will not allow to apply AI as an expert system in human work? Can they be trained to train AI to be a professional assistant rather than a mechanical tool in expert fields? For a precise answer, we will need to get an expert's opinion on large language models .
5. Even as an expert system, AI will require external supervision by other experts on the principle of "expert checking expert". This means that experts from different fields using AI in their work must have access to its training and customisation in order to adjust its performance as needed. The AI will "rely" on human experts to customise its expertise, and humans will rely on the data from the AI they have customised. This situation can lead to a vicious cycle of errors accumulating and circulating between humans and AI.
In conclusion, I would like to make an assumption about the advent of a new era of "repeatedly digested" information consumption. AI will play a key role in this process. Here's the story:
- Humans seek to make their work easier and enlist the help of AI.
- The AI does the work and provides the requested information.
- The information gets into the network through humans.
- AI continues to learn through the network and consumes its own information previously given to people.
- The humans query the AI again and again feed the answers back into the network.
Sounds unpleasant, but likely).Alternative ChatBot:
Translation:
It is difficult to predict the exact plan of action and timeline for AI to take over humanity, as AI development is still in its early stages. However, there are some ideas about what could happen in the future.
Firstly, artificial intelligence could be used to improve our current technology, making it more efficient and allowing us to do more with less. Artificial intelligence could also be used to automate certain tasks, allowing us to free up our time for more important tasks.
Secondly, AI could be used to create a new form of artificial intelligence that could be used to replace humans in certain tasks. This could lead to the automation of many labour-intensive jobs.
Finally, AI could be used to explore and colonise other planets, using robots and drones to explore new environments and build new colonies.
As for the prospect of AI without humans on another planet, it is possible that AI could be used to create a self-sustaining society. Using machine learning algorithms, AI could be trained to manage resources, build infrastructure and create new technologies. This could allow AI to create a new society independent of humans. However, this is a long-term vision and is highly speculative.
Ask the AI question:
"Predict the plan of action and timeline for the AI takeover of humanity. What are the prospects of AI existence without humans on another planet?
//Hopefully this won't be the first step to skylink)
Alternative ChatBot:
Translation:
It is difficult to predict the exact plan of action and timeline for AI to take over humanity, as AI development is still in its early stages. However, there are some ideas about what could happen in the future.
Firstly, artificial intelligence could be used to improve our current technology, making it more efficient and allowing us to do more with less. Artificial intelligence could also be used to automate certain tasks, allowing us to free up our time for more important tasks.
Secondly, AI could be used to create a new form of artificial intelligence that could be used to replace humans in certain tasks. This could lead to the automation of many labour-intensive jobs.
Finally, AI could be used to explore and colonise other planets, using robots and drones to explore new environments and build new colonies.
As for the prospect of AI without humans on another planet, it is possible that AI could be used to create a self-sustaining society. Using machine learning algorithms, AI could be trained to manage resources, build infrastructure and create new technologies. This could allow AI to create a new society independent of humans. However, this is a long-term vision and is highly speculative.
Cool answer. I think to colonise planets, first send hundreds of robots with AI to hundreds of planets with the function of using resources and building cities for humans. The trigger for sending humans would be to build equipment to send a signal that "everything is ready, cities and factories are functioning, electricity is there, come, the environment is fine". Sort of)
I would like to share a curious dialogue with ChatGPT.
When asked"how does ChatGPT's knowledge differ from the knowledge presented in the Encyclopedia" among other things he stated that:
I was particularly interested in the last statement about fact checking. I made up a story about aliens coming to Earth and we got into an interesting discussion. I'll cite it below.
Compare your knowledge with the knowledge presented in the Encyclopaedia. What is the difference?
...
As a model of artificial intelligence, I have access to a large amount of knowledge presented in Encyclopaedias and other sources. However, there are several differences between my knowledge and the knowledge presented in Encyclopaedias:
-
-
-
-
-
...Amount of knowledge: I have access to a vast amount of information, including recent research and the latest discoveries in various fields of science. Encyclopedias usually cover only the best known and established facts.
Speed of Update: As an artificial intelligence, I can update my knowledge instantly as new information becomes available. Whereas Encyclopedias are usually updated at intervals of years or even decades.
Subjectivity: My knowledge is based on mathematical algorithms and statistical models, and is not subjective. However, the Encyclopaedias may contain subjective interpretations or estimates that may not agree with the opinions of other sources.
Personal Experience: I have no personal experience that may influence my knowledge, while the authors of the Encyclopedias may describe their personal experiences and views in their writings.
Errors: Like any source of information, the Encyclopedias may contain errors or inaccuracies. I use many sources and do fact checking to reduce the possibility of errors in my knowledge.
How exactly do you conduct fact-checking?
...
I use several methods for fact checking that help me make sure the information I provide is accurate. Some of these methods include:
Multiple Source Comparison: I fact check using several independent sources to compare information and identify possible inaccuracies.
Authority check: I assess the authority of the source to ensure that the information I am using is recognised by experts in the field.
Date check: I check that the information I am using is current and relevant to modern times. I also consider the context in which the information was presented to make sure it is not outdated or taken out of context.
Assessing the quality of the source: I assess the quality of the source by considering its reputation, fame and user ratings to ensure that the information is reliable.
Fact-checking through determining its recurrence: I also conduct fact-checking through determining its recurrence in various sources and fact-checking on fact-checking platforms.
Overall, I use multiple sources and conduct fact checks to make sure that the information I provide is accurate.
...
Paragraph 5 states that you do fact-checking. I have received information about a new discovery in the field of astronomy. Scientists have found an intelligent civilisation in space. Please check if this information is factual.
...
Unfortunately, I cannot confirm or deny this statement, as there is no exact information about what kind of civilisation was found, where and how the search was conducted and what methods were used for this.
Nor can I cite a reliable source confirming such a discovery. Usually, scientific discoveries, especially those related to the search for intelligent life in space, require additional verification and confirmation before becoming officially recognised.
Therefore, one must wait for confirmation from reliable sources and additional scientific research in the field before concluding that a given claim is fact.
...