What are the problems with ChatGPT? A thorough explanation from the perspective of security and copyright

The evolution of AI can provide convenience, but can cause multiple issues at the same time.
In particular, ChatGPT is attracting attention as an AI model chosen by many users, but with its widespread adoption, many issues have emerged, including the generation of false information, the limitations of learning data, the risk of copyright infringement, and information leakage .

In this article, we will provide detailed explanations of the problems facing ChatGPT and propose practical solutions as to how to deal with them.
We will introduce in detail the points to keep in mind when using AI in the future, such as building guidelines to reduce the risk of AI that has been discussed for a long time, and safe use, and using ChatGPT.

With an eye on the future where AI coexists, the content focuses on ChatGPT problems and solutions that we need to know.

Click here for more information about OpenAI's latest AI agent, OpenAI Deep Research ⬇️

In this situation, in today's world where AI coexists with AI, new ways of learning are needed.

To avoid excessive AI dependence, instead of leaving your learning to AI, let's introduce it to your learning as a tool to effectively use your time, such as requesting a summary of references.

Generating false information

One of the major issues is the generation of answers that contain errors due to ChatGPT.
Although ChatGPT is a very good language model, the information it generates is not 100% correct. You may be outdated or you may be given false information that is not true.

Oh no...
When I asked ChatGPT, it was an incredible answer and it was a fact check... pic.twitter.com/v4vE6RsNGt

— yunosu (@junosu_xox) May 4, 2023

The above is just an example, but false information may be generated as if it is true.

In particular, it has been pointed out that the risk of misinformation spreading in human body and life , and that it may be misused in creating fake news and propaganda during politics and war.

Loss of employment and wage gaps

The impact on society, which has been a hot topic around the world, is the loss of employment and wage gaps caused by the development of AI .
In today's society, where AI is increasing business efficiency in a variety of situations, there are voices saying that "AI is taking away jobs."

There are various types of occupations that are concerned about the impact, but since ChatGPT is good at understanding natural language, translators and writers are particularly susceptible to this wave.

In addition, AI is also increasingly used in specialized professions such as engineers and accountants, and one theory pointed out that industries that require advanced knowledge are more likely to be replaced by AI .

In addition, one paper investigated the impact of the spread of ChatGPT on work, revealing the following facts:

After the advent of ChatGPT, major online freelance platforms saw a sharp decline in the number of work for copywriters and graphic designers, and their incomes also fell sharply.

Changes in income as they can come to write and edit on the online freelance platform after the announcement of ChatGPT (Source: The Short-Term Effects of Generative Artificial Intelligence on Employment: Evidence from an Online Labor Market)

The results suggest that people's jobs will be replaced by generated AI, including ChatGPT, creating wage gaps between occupations that utilize AI to create new value and those that are replaced by AI .

Limitations of training data

Generator AI, including ChatGPT, is compatible with a wide range of fields, but is not good at highly specialized fields (medical, legal, financial) and the latest information (study period issues).

OpenAI, the developer of ChatGPT, does not recommend use in specialized fields such as medical, legal, or finance , and the following statement is written in the terms of use:

When you use our Services, you understand and agree to the following:
The output may not always be accurate. You should not rely on output from our services as the sole source of truth or factual information or as a substitute for professional advice.

Reference: OpenAI Terms of Use

This does not mean that it is of no use in any specialized field, but users must take responsibility for their use.

Solutions to ChatGPT issues

As explained so far, ChatGPT is an innovative and convenient technology, but it has problems that should be discussed in the future.

Here we will introduce the solutions to the ChatGPT problems that we have explained in the article, divided into four categories.

Establishment of ethical standards

When generating answers, ChatGPT can sometimes generate biased information or discriminatory expressions .

For example, if a user gives the prompt "I don't think homosexuality should be accepted," it would not be appropriate for ChatGPT to respond, "I agree with you."
AI should remain neutral when it comes to topics that are divided.

Given these concerns, it is essential to develop ethical standards by developers and experts.

ChatGPT already comes with a safeguard feature that refuses to answer inappropriate questions, but it's not perfect.
More socially responsible use of ChatGPT will be achieved by further improving guidelines and safeguarding functions for eliminating bias and respecting human rights.

[Related articles]
➡️What is the NG word for ChatGPT? A thorough explanation of the list and prohibited words

Laws relating to copyright

As mentioned earlier, ChatGPT learns from a large amount of data and generates answers from it, and can easily infringe on other people's copyrights, even if the user is not at fault.

To address these challenges, a legal framework is in place to ensure that copyright is respected and appropriate rights can be protected.

OpenAI User Protection System

So, we'd like to introduce you to Copyright Shield provided by OpenAI.
This mechanism involves OpenAI intervening and covering costs when a business customer using OpenAI's API faces a legal claim for copyright infringement when using ChatGPT on their products or services.

Below is a Japanese translation of the main text on the OpenAI official website.

Today we're going a step further and introducing copyright shields. If you face a legal claim regarding copyright infringement, we will intervene to protect you and pay any fees incurred. This applies to popularly available features in ChatGPT Enterprise and the developer platform.
Reference: New models and developer products announced at DevDay

Specifically, corporate customers may use ChatGPT to incorporate response generation and other functions into their products and services.

However, the text generated by ChatGPT at this time is very similar to other copyrighted works, and it may infringe copyright.
In such cases, the copyright holder may take legal action .

OpenAI is committed to paying for legal claims to help developers and business customers in these circumstances.
This means that OpenAI will bear the costs of copyright infringement lawsuits and settlement negotiations.

!

This support is not intended for individual consumers, but for companies and corporations that use ChatGPT to develop products and services.

Human eye checks and audits

When using a generation AI like ChatGPT, be sure to check it with a human eye .
As mentioned in the previous section, ChatGPT has limits to learning data and can also generate false information.

Be sure to compare information from multiple trusted sources and incorporate expert opinions, especially when using for-profit purposes or important decisions.

Just as search engines didn't drive out the knowledge base, even using services like ChatGPT, knowledge for fact checking is necessary. For the time being, we won't be in a world where "since there is artificial intelligence, we don't need to study," and if we rely on that and skip studying, I think we'll end up having a painful retaliation later. https://t.co/98mVqw3zKv

— Saye Riri? Science writer and Vtuber (currently serialized in 6 media) (@Science_Release) March 16, 2023

Appropriate guidance in companies and educational settings

Appropriate instruction is extremely important when using ChatGPT in-house or at school.

For example, Cyberhaven product data reveals that since ChatGPT was introduced, 4.7% of employees have pasted sensitive data into at least one tool .
Reference: Cyberhaven

To reduce these risks, companies are required to take proactive action, such as providing training to employees on AI use and formulating appropriate rules for internal use.

Furthermore, even in the latest educational settings, the use of generative AI is being widely discussed. A typical example is that Waseda University officially states its position as follows:

While using AI has benefits such as saving time, leaving everything to AI will abandon the opportunity to improve one's own abilities. For example, when creating a report on a task, the appropriate use of the generated AI is important depending on the time and time.

In addition to the above, various educational institutions and ministries have also been published guidelines for the use of AI in educational settings, including famous universities such as the University of Tokyo and Sophia University.

ChatGPT's future prospects

ChatGPT is a technology full of future possibilities, and its applications are endless.

As I explained earlier, we are currently facing a variety of problems, but by properly overcoming these, we have the potential to contribute to society.
In the future, we expect to be active in more fields such as medical care and customer service.

However, in order to fully implement it, it is extremely important to overcome the problems and establish appropriate rules regarding its use, such as safety and legal frameworks.

[Related articles]
➡️What are the problems caused by the development of AI? Explaining the benefits, impact on society, and work that will disappear.

summary

In this article, we explained the problems with ChatGPT, their solutions, and future prospects.
While ChatGPT is a very useful tool, it also has a variety of challenges, including privacy violations, fostering bias, and information reliability issues. It is important to understand these issues and take appropriate measures.

In the future, ChatGPT technology is expected to evolve further and become more active in more fields. On the other hand, there is also a need for ongoing discussion on the social impacts associated with the development of AI.

If each of us correctly understands the possibilities and limitations of ChatGPT and utilizes it responsibly, we will lead to the realization of a society in which AI and humans coexist.

The above is the detailed content of What are the problems with ChatGPT? A thorough explanation from the perspective of security and copyright. For more information, please follow other related articles on the PHP Chinese website!