Ce que ChatGPT peut apporter en matière d'apprentissage et d'efficacité

 De : https://articles.mercola.com/sites/articles/archive/2024/02/22/transformative-potential-chatgpt.aspx

Traduction automatique ici : http://tinyurl.com/4u3tayxh

The Transformative Potential of ChatGPT in Learning and Efficiency

Analysis by Dr. Joseph Mercola  22 février 2024

• 
  

Story at-a-glance

• The fear of new technologies, be it calculators or ChatGPT, often stems from a lack of understanding and apprehension in the face of change. In the case of ChatGPT, we are witnessing a phase where initial fears are predominantly based on misconceptions about AI taking over human tasks entirely
• While a 2013 study by the University of Oxford predicted that 47% of the American workforce might be replaced by AI over the next two decades, that prediction is unlikely to come to pass. Because AI hallucinations are a permanent feature, the need for human judgment will remain whenever and wherever AI is used
• ChatGPT is designed to be a tool that augments human capabilities. It can facilitate various tasks, like generating ideas, composing texts, coding and more
• ChatGPT is akin to a pipeline, channeling vast amounts of information and insights to disseminate knowledge more broadly. This efficiency is not about replacing human action but enhancing it, allowing individuals to focus on aspects of their work that require uniquely human attributes like empathy, strategic thinking and creativity
• Six key precautions when using ChatGPT and other generative AIs are discussed

In an era where technological advancements are revolutionizing every aspect of our lives, ChatGPT emerges as a beacon of innovation, especially in the realms of learning and efficiency.

I was really surprised that 61% of you disagreed with me on the value of ChatGPT, even though I believe the article I wrote was well balanced and had an accurate description of the concerns and strategies to mitigate these concerns. But most of you did not see it that way.

Many may not remember that technology has been a lifelong passion of mine and I am an early adopter. I took my first computer programming class in 1968 and was first on the internet in 1978, long before the web began. I started my website in 1997, before Google even existed. I recognized that the internet was the greatest innovation in history at that time, which is why I started my website so early.

I grabbed a ChatGPT account on the first day it was available in November 2023, because I recognized that it eclipsed the internet in terms of its revolutionary impact. Many of you may need a few years to catch up, but eventually you will recognize this.

However, its introduction is not without controversy, much like any groundbreaking technology that has preceded it. This controversy stems from a fundamental misunderstanding of the role of such tools in augmenting human potential, rather than replacing it. This realization really hit home when I wrote my article on this topic last month.

At the heart of this misunderstanding is a fear reminiscent of the anxieties stirred a century ago with the advent of automobiles, replacing horse-drawn carriages. The automobile revolution brought about significant cultural and economic shifts, some of which were feared and resisted.

Overcoming Fear in Enhancing Human Capabilities

The fear of new technologies, be it calculators or ChatGPT, often stems from a lack of understanding and apprehension in the face of change.

When calculators were first introduced, there was concern about their impact on traditional calculation methods and the potential dependency on this new tool. Similarly, the initial apprehension toward ChatGPT and similar AI technologies revolves around the fear of over-reliance and the potential loss of certain skills or jobs.

With calculators, people eventually realized that these devices were not here to replace human intelligence but to enhance and assist in solving complex problems more efficiently. Calculators became indispensable tools, widely accepted, and integrated into education and professional settings. The fear subsided as the benefits became clear and the tool was understood as a complement to human capabilities, not a replacement.

To illustrate this further, consider the modern spreadsheet, a significant technological leap following the calculator. This tool efficiently processes the data you input, transforming it into a format that's much clearer and simpler to understand.

Essentially, the spreadsheet software acts as an extension of your brain power, simplifying data interpretation and enhancing your understanding of what the data signifies. It's widely acknowledged that spreadsheets are a helpful tool, and there's generally no fear associated with using them.

In the case of ChatGPT, we are witnessing a similar phase where initial fears are predominantly based on misconceptions about AI taking over human tasks entirely. However, just like calculators, ChatGPT is designed to be a tool that augments human capabilities. It assists in various tasks, like generating ideas, composing texts, coding and more, thus enhancing our productivity, creativity and joy.

The comparison highlights a recurring pattern in the adoption of new technologies. Initial fears and resistance give way to acceptance and integration as the real value and role of technology as an aid rather than a replacement becomes evident.

Over time, just as the fear of calculators subsided and they became an integral part of our toolkit, the apprehensions surrounding ChatGPT and similar AI technologies are likely to diminish as their role as valuable tools in extending human capability is more widely recognized and understood.

Enhancing Human Potential, Not Replacing It

Another fear surrounding ChatGPT revolves around the concern of human redundancy — the fear that machines capable of mimicking human thought processes might render human cognition and action unnecessary. This fear, while understandable, overlooks the essence of what tools like ChatGPT are designed to achieve.

They are not replacements for human intellect or creativity, but facilitators that amplify these intrinsic human qualities. ChatGPT is akin to a pipeline, channeling vast amounts of information and insights, not to concentrate power or resources, but to disseminate knowledge more broadly.

This efficiency is not about replacing human action but enhancing it, allowing individuals to focus on aspects of their work that require uniquely human attributes like empathy, strategic thinking and creativity.

This dissemination of knowledge is particularly crucial in the field of education and learning. ChatGPT, with its ability to understand and generate human level insights, can be an invaluable tool for students and educators alike. I know that I have benefited greatly from its use as a tool to help me understand important concepts. It is exponentially superior to any search engine I have ever used.

It can provide instant access to information, aid in the composition of essays and reports, assist in language learning, and offer explanations of complex concepts, especially in health, biology and medicine. This capability enhances your learning experience, making it more interactive, personalized and accessible.

Download this Article Before it Disappears

Download PDF

Maximize Your Efficiency and Achieve Your Goals

ChatGPT can radically improve your life if you use it carefully. My last article provided all the necessary cautions, some of which are summarized below, that you need to consider when using this tool. That said, let me give you some examples of what you can use ChatGPT for:

• What do you want to do at this moment?
• What do you want to learn?
• What do you want to focus on?
• What action do you want to take?
• What do you want to change?

All those things are choices in the moment, and that is what ChatGPT facilitates. It provides you with radically more efficient information gathering in the moment, and a drastic reduction in wasted time and effort in finding information that would allow you to achieve your goals.

There are many of you who are unable to understand the smooth transition of your thought to action, or intent to action, through this technology. Rather, you think that you are no longer required, because you are not well-connected to your consciousness and your own sense of self.

You don’t yet realize that this is only a tool to facilitate and improve who you are. You see yourself as limited to your biology, and that is just not accurate. By doing so, you are limiting your understanding of your reality.

Consider ChatGPT as a Complement to Your Consciousness and Creativity

The effectiveness of ChatGPT in learning and efficiency hinges on your perspective. If you view it merely as an extension of your own mind, its potential is significantly limited. In fact, this view of technology radically increases your fear of it. This is because many people don't understand that their mind and five senses have relatively limited data, compared to the ocean of data in their consciousness.

Your mind can be compared to a thimble that has a limited capacity to store information compared to the vast ocean of knowledge that is available outside our five senses in the realm of consciousness.

ChatGPT can be used as a tool that complements and extends your own “ocean” of knowledge and creativity if its true value is understood and realized. It can then become an ally in the learning process, and an aid that fosters deeper understanding, stimulates curiosity, and encourages independent thought.

By embracing this technology and integrating it into your workflows and learning processes, you can expand your horizons, enhance your capabilities, and journey toward a future where technology and human potential synergize to create a more informed, efficient and creative world.

You just need to be careful, as I outline in the precautions below. I understand the biases that are currently programmed into this technology, but this can change. In fact, there is a great likelihood that I will be meeting with Sam Altman, the founder and current CEO of open AI, later this year to discuss and facilitate an option to remove this bias.

In addition to the six key precautions below I strongly encourage you to consider before using ChatGPT, the most crucial is to never forsake your critical thinking skills while using it. This is arguably its greatest limitation.

ChatGPT is not a replacement for your own critical thinking; rather, it serves as an aid to enhance your creativity. It's vital to keep this in mind and remember that ChatGPT is an adjunct, not a substitute, for your intellectual and creative capacities.

Six Key Precautions

Some of the key precautions to keep in mind when using ChatGPT (or similar large language models) include the following:

1. Protect your privacy; never share personal or confidential information about yourself or others — ChatGPT stores every conversation you have with it on OpenAI’s servers, and these logs are shared with other AI companies and AI trainers. As a result, any private information entered into chat can end up in responses to other people’s questions.

2. Be aware that ChatGPT gathers things like your IP address, browser type and browser settings, your interactions with the site and your online browsing history, and that OpenAI may share all of this information with unspecified third parties¹ — You consent to all that data gathering and sharing when you accept OpenAI’s privacy policy.²

3. Log in using a throwaway email address — ChatGPT also stores the email address you use when logging in, and in a recent experiment, Ph.D. candidate Rui Zhu was able to extract the email addresses of New York Times personnel and other users.³

The experiment has revealed a security loophole that could be exploited by malicious actors for phishing purposes and the like, so, to be on the safe side, you may want to use a throwaway email address when logging in rather than your personal or work-related email.

4. Find out your employer’s use policy — Due to the privacy risks, some companies and government agencies are prohibiting the use of ChatGPT on its networks, so make sure you know your employer’s policy.

For example, in October 2023, the U.S. Department of Agriculture (USDA) banned the use of third-party generative AI tools by employees and contractors in their official capacities and on government equipment due to data safety risks.⁴

In another example, staff at NASA are only permitted to use ChatGPT on personal accounts and must follow “acceptable use policies.” A helpful list of tips to prevent security risks from generative AI within an organization can be found on IANS Research blog.⁵

-5. Be mindful that ChatGPT may hallucinate (fabricate information) and can offer persuasive yet factually incorrect arguments — To minimize this risk, ask it to provide source links and verify the accuracy of the information given. Also, do not rely on ChatGPT-generated information in cases where doing so might have catastrophic consequences.

6. Keep in mind that there are many open legal questions when it comes to who owns the rights to ChatGPT-created images, content and code⁶ — As noted in an IBM blog post,⁷ “Terms of service state that the output belongs to the provider of the input, but issues may arise when the output includes legally protected data sourced from other inputs. Copyright concerns⁸ may also arise if ChatGPT is used to generate written material based on copyrighted property.”

This is yet another reason why you’d want to use your own discernment and double-check everything. As just one example, if you’re using ChatGPT to produce written material, you need to make sure nothing has been plagiarized (taken word for word from a source without proper attribution).

Time will tell how that all pans out. In the meantime, use ChatGPT-generated output with caution, lest you get caught up in some copyright or license infringement dragnet later on. Getty Images has already raised a legal challenge⁹ against Stability, as its Stable Diffusion image generation tool was trained on Getty images and are now being used to create new images without paying licensing fees.

Replacement Theory May Be Off-Base

As reported by Business Insider¹⁰ in February 2023, while a study¹¹ by the University of Oxford, published in 2013, predicted that 47% of the American workforce might be replaced by AI over the next two decades, that prediction “appears to have been off-base ... because human judgement still needs to be applied to these technologies to avoid error and bias.”

Anu Madgavkar, a partner at the McKinsey Global Institute, told Business Insider,¹² “We have to think about these things as productivity enhancing tools, as opposed to complete replacements,” which is precisely the stance I’m advocating for.

As noted by IANS Research,¹³ AI hallucinations “are not going away. They are a feature, not a bug.” So, human discernment and expertise will remain a crucial component whenever and wherever AI technology is used.

 Sources and References

• ¹ The Conversation February 8, 2023
• ² OpenAI Privacy Policy. January 31, 2024
• ³ Business Today December 26, 2023
• ⁴ Fedscoop December 20, 2023
• ^{5, 6, 9, 13} IANS Research June 27, 2023
• ⁷ IBM.com August 16, 2023
• ⁸ The Verge, The Lawsuit That Could Rewrite the Rules of AI Copyright November 9, 2022
• ^10, 12 Business Insider February 2, 2023
• ¹¹ Working Paper The Future of Employment September 17, 2013 (Archived)