Artificial intelligence, or AI, has come a long way in just the past few months. As we learned in last week’s post, AI really can be a helpful tool, but you must consider its limitations when deciding whether it’s worth using for your specific purposes.

This week, we will look at the next three questions we asked AI to learn more about the data it uses to generate text for you. TextRanch editors reviewed each of the answers AI wrote to identify any English errors or other problems. You can find the editor’s notes and suggestions under each of the answers below.

Question 3*: “Is my information used to train AI?”

Here is AI’s response:

Yes, personal data is commonly used to train AI models, it is used to learn patterns and relationships between inputs and outputs, so the AI can perform tasks and make decisions based on the data it has been trained on.

Editor comments:

There are several errors in the sentence structure. The phrasing in the last part of the sentence is vague and does not add much value to the answer.

Specific issues with this answer:

“…used to train AI models, it is used…”

Problem: This comma splice (see the explanation about comma splices in last week’s post) is creating a run-on sentence. A run-on sentence happens when two independent clauses are joined incorrectly.

Solution: A run-on sentence can be fixed by creating separate sentences. Do this by replacing the comma with a period and starting the new sentence with a capital letter.

“…inputs and outputs, so the AI can…”

Problem: Although the comma and “so” are used correctly, the larger sentence is still too wordy.

Solution: Replace the comma with a period and start a new sentence with a capital letter. A human editor will need to rephrase the new sentence because it is better not to start a sentence with “So”.

“…the data it has been trained on.…”

Problem: It is okay to end a sentence with a preposition (“on”) in some situations. However, it sounds awkward here.

Solution: A human will need to rephrase the words to form a full, complete sentence that does not end with a preposition. (For more information on how to fix an incomplete sentence, check out this previous series of articles.)

Corrected version:

Personal data is commonly used to help AI learn patterns and relationships between inputs and outputs. AI requires new data to keep learning and improving its ability to perform various tasks. When you enter any data into an AI tool, it could be used to help train that tool.

Question 4*: “Can I trust that my information will be kept private if I use AI?”

Here is AI’s response:

The trustworthiness of keeping your information private when using AI depends on the specific implementation and usage of the AI system, additional steps must be taken to ensure data remains private, such as data privacy, data anonymization, and differential privacy, and ensure that the data is collected, stored, and used in a secure way.

Editor comments:

There are issues with sentence structure as well as vague wording.

Specific issues with this text:

“…usage of the AI system, additional steps must be taken…”

Problem: Like the previous answer, this one includes a comma splice that creates a run-on sentence.

Solution: Put a period at the end. Start a new sentence with a capital letter.

“…data remains private, such as data privacy,…”

Problem: The wording is awkward. Specifically, it is missing a verb.

Solution: Put a period after “private”. A person can rewrite the next segment as a complete sentence that includes a verb. (For more information on how to fix a run-on sentence that is missing a verb, click here.)

“…, and ensure that the data is collected…”

Problem: It is better not to start a sentence with a conjunction such as “And”.

Solution: A human writer can rephrase the sentence so it does not start with “And”.

Corrected version:

The trustworthiness of keeping your information private when using AI depends on the specific implementation and usage of the AI system. Additional steps must be taken to ensure that data remains private. Some examples include using data privacy, data anonymization, and differential privacy. This will ensure that the data is collected, stored, and used in a secure way.

Question #5*: “Can I trust AI to produce writing that is free of plagiarism?”

Here is AI’s response:

AI-generated content is likely to be plagiarism-free, but it’s recommended to run it through a plagiarism checker before publishing. Additionally, it’s important to make sure that the AI model is not trained on copyrighted or proprietary data, as it could generate plagiarized text.

This answer is grammatically correct and doesn’t need any changes.

What does it all mean?

Based on AI’s answers, it would be a smart move for you to find out more about the exact procedures and policies of any AI tool that you are using. Just like with any other online tools you already use, AI has similar risks and benefits.

In terms of data sharing and privacy, it is nearly impossible to lead a totally private life while using Google, Facebook, Amazon, or any other major online platform. This is the price we all pay for their “free” services, and many AI tools will work the same way.

Plagiarism and copyright issues are more difficult for AI users to monitor. Beyond being respectful of the original writers, for many types of writing (e.g., schoolwork, anything you will post publicly, any commercial writing) you must be aware of the risks of using someone else’s work. Pairing AI with a quality plagiarism detector can help protect you from copyright infringement.

AI can be very helpful if you think of it as just one of many items in a writer’s toolkit. To achieve the best results, you will still need human editors, a plagiarism checking tool, and a fact checking tool. To get the best responses for your particular situation, you may also need to spend extra time testing various prompts for the AI tool.

In our next article, we will take a closer look at the error we most frequently identified in AI content, the comma splice/run-on sentence issue. This is a common writing problem, and evidently AI has picked up this habit from its human data donors. Therefore, we will examine this issue in greater depth.