[Caution] Could your ChatGPT conversations have been seen by someone on Google? The hidden side of AI reported by overseas media

Do you really believe your conversations with AI are a "secret between just the two of you"?

Have you been using AI chatbots like ChatGPT lately? I have also been relying on them more and more for gathering information and writing articles. It feels just like learning the "Zoom" spell in Dragon Quest—the sensation of reaching your destination in an instant. It really is convenient, isn't it?

But what if there is a story hidden behind that overwhelming convenience that we should stop and think about for a moment...?

I recently found an article on the famous overseas site Lifehacker that made me exclaim, "Wait, really?"

The title is "You Might Have Sent Your ChatGPT Conversations to Google".

Translated into Japanese, it means something like "Your conversations with ChatGPT might have been sent to Google."

This article is just one report from an overseas media outlet, but if this were true, wouldn't it be a little chilling?

Many people believe that their conversations with AI are completely private. I decided to cover this article because I felt that by shedding light on this "unknown possibility," it might serve as an opportunity for everyone to re-examine their own digital lives.

Just knowing that "this perspective exists" might change how you interact with AI in the future.

I would be happy if you could stay with me until the end.

Reference: Lifehacker "You Might Have Sent Your ChatGPT Conversations to Google" (https://lifehacker.com/tech/you-might-have-sent-your-chatgpt-conversations-to-google)

Are AI conversations leaking? The shocking truth reported by overseas media

According to the Lifehacker article, OpenAI, the company that operates ChatGPT, was temporarily conducting an experiment that made some user conversations discoverable via Google search.

You might think, "How could that be possible?" ChatGPT is like your own secret confidant. Some people might even be sharing things they wouldn't tell anyone else, from work troubles to personal matters.

The possibility that this was in a state where anyone could see it via Google search is unsettling, even just to imagine.

According to the article, there were cases where records of conversations users had with the AI were actually found using specific search methods.

Some of these reportedly included content that could identify individuals from resume consultations, as well as highly sensitive conversations. While this is only what is reported in the article, if you think about your own diary being published without your knowledge, you can understand the magnitude of the impact.

Why conduct such an experiment? Deciphering the intentions of AI companies from the article

So, assuming this experiment was taking place, what was OpenAI's intention?

The comment from OpenAI introduced in the article was that they were 'testing ways to make it easier for users to share helpful conversations'.

From here on, this includes my personal speculation, but I feel that behind these words lies a major trend in the entire AI industry.

The author of this article, Jake Peterson, is a journalist who has been covering the technology sector for many years.

Perhaps he is sounding the alarm on this issue because of his perspective as an expert that we should look at the potential risks of AI evolution alongside its benefits.

Thinking based on his article, the possibility emerges that OpenAI's goal was to 'show many people excellent AI-generated answers and increase their influence in the search market'.

One could also view it as an attempt to increase the value of ChatGPT itself and establish its position as a source of information by displaying user conversations in search results.

This could be seen as a sign that AI is evolving from a mere tool into an 'information platform'. Of course, this is just one interpretation I have read into the content of the article.

Are your conversations 'training data'? AI mechanisms you should know about

The most important point we should learn from this incident may be to have the awareness that conversations with chatbots are not necessarily 'private'.

OpenAI states that it has ended the experiment in question, but a more fundamental issue remains. As emphasized in the article, many AI companies use conversations with users as 'training data' to improve AI performance. Furthermore, human reviewers may also view those conversations for quality control purposes.

This could be compared to a situation where a secret you confided in someone is accumulated as that person's knowledge and potentially shared with someone else.

Although there is a setting in ChatGPT to prevent conversations from being used as training data, it is said that the records of the conversations themselves are still stored on the server.

Even 'temporary chats' are said to remain for up to 30 days. It's like thinking you used 'vanishing magic' when, in fact, the traces were left behind quite clearly.

Perhaps we need to be 'heroes' who master powerful AI tools, while also being 'sages' who calmly assess their mechanisms and risks.

Equip yourself with 'digital armor' to live wisely in the AI era

From this Lifehacker article, we have gained one important perspective: 'conversations with AI may not be as private as you think'.

This does not mean we should blindly fear AI. I see this as a good opportunity to think about how to interact with it wisely, once we understand its characteristics.

For that reason, isn't it important for us to be conscious of equipping ourselves with 'digital armor'?

Practical advice for protecting your digital life

Based on the content of this article, I would like to propose a few self-defense measures that I have thought of. I leave it to your judgment to decide what to do in the end.

1. Be careful when entering confidential information
One option is to decide not to enter personal information, company secrets, passwords, or any other information that would be problematic if leaked into a chatbot.

2. Try checking your privacy settings once
I recommend reviewing the privacy settings of the AI services you use. Just knowing how your data is handled should change your awareness.

3. Don't over-rely on 'temporary conversations'
It might be good to keep in the back of your mind that even in modes that don't save history, there is a possibility that data remains on the server.

4. Keep your antenna up for information regarding AI
Information regarding the mechanisms and risks of AI is updated daily. Continuing to learn from reliable sources might be the most powerful 'armor' of all.

The future of coexisting with AI has already begun. If this story serves as a 'trigger' for you to re-examine your own digital life and engage with AI more safely, there would be nothing that makes me happier.

If you liked this, please hit the heart button and follow me! It really motivates me!