Recently, my friend and I talked about his commonly used AI input method, sharing some of the recent AI input methods that have brought great convenience to daily life and work.

His sharing immediately triggered some sensitive nerves in me regarding input methods and AI.

In the past, I had used a certain domestic input method for a while. It indeed brought me great convenience in inputting and improved my efficiency. However, after using it for some time, I gradually found that some ads related to my input habits would pop up from time to time. So I quickly uninstalled this input method and reverted to the most basic and "clumsy" input method that came with the operating system.

Since then, I have hardly been bothered by new ads, and I have also heightened my attention to and protection of privacy.

I have seen some introductions to the recent AI input methods on various channels, but because of my past experience, I have always had reservations about privacy breaches and therefore have not tried these new tools.

At the same time, I have felt a greater risk of personal privacy breaches in the AI era in other contexts.

Some readers have left comments asking me what AI tools I commonly use. The main one I use is Google's Gemini, followed by Tencent's Yuanbao, and occasionally I use Manus.

I use Google's Gemini almost every day, and I use it heavily. Basically, any question I think of, I will communicate with it immediately.

Before using any AI tool (including Gemini), we are actually very clear that all the data we provide to AI will be used for training its corpus.

But what consequences this training will bring, whether it will affect us or even one day backfire on us, I guess the vast majority of users may not have a strong sense of it right now.

For example, Gemini has always provided me with a very good user experience, and I have never thought about any "backfire." But two incidents that happened last week triggered my more sensitive nerves.

When I was discussing a question about nutrition with Gemini, it suddenly said (in essence): the situation regarding your mother that you asked about some time ago falls under this case.

I was a bit surprised when I read this sentence.

Because when I previously asked it about my mother, I never said it was my mother, so I was very curious how it guessed that it was my mother?

Next, it provided a solution and explained that this solution is applicable not only to your mother but also to you.

This made me even more curious: does this mean it has grasped the issues, health status, and potential consumption trends regarding diet and nutrition for both of us?

The second incident was when I asked it some questions about the A-shares, it suddenly popped out an overall evaluation, commenting on the risk relationship between the crypto assets, US stocks, and A-share assets that I am concerned about, and also roughly described my risk preference and asset allocation.

Looking back at the various investment-related questions I asked it in the past, they were actually quite scattered, simply whatever I felt like asking. I never told it how much I invested or what specific assets I had invested in.

But I estimate that it may have gradually outlined my style, risk preference, and even guessed the proportions of my various assets based on the frequency of my questions and focal points of concern.

More importantly, using Gemini requires logging in with an email account, which contains various email information about me.

If Gemini aggregates my email information and the various questions I have discussed, then aside from not knowing my real name and address, it can almost guess every aspect of my life - from daily habits to health status.

This kind of invasion of privacy has reduced the defenses against personal privacy infringements from current internet giants.

If this privacy data is used illegally, it can completely disrupt a person's life or even determine a person's life or death.

Among many internet giants, Google at least has a certain bottom line in terms of respect for personal privacy. What would happen if we encountered those AI tools that have no bottom line, and our personal privacy was collected by them?

This is simply unimaginable.

After these two incidents, I am now particularly cautious when using AI tools.

Firstly, I will try to use different tools in conjunction to avoid concentrating personal information too much on a single tool.

Secondly, when selecting AI tools, I will pay more attention to the values of the AI company, seeing if it has any bottom line and whether it can have at least some basic respect for user privacy.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。