人工智能正在让你变得无聊

Speaking of AI helping us make decisions, I would like to ask everyone to imagine a scene first.

It is a Friday evening, you have worked for a week, and you feel very tired.

You want to go out for a nice meal, but you don't want to spend time looking for a restaurant.

So, you take out your phone and ask your AI assistant to help you book a restaurant.

You tell it that any restaurant with seats on Friday night and good reviews will do.

The AI starts working immediately, and its speed is very fast.

Within one second, it checks all the reviews on the internet.

It compares the scores of different restaurants and looks at many photos.

Finally, it books a very famous restaurant in this city for you.

This restaurant is a place many of your friends have been to.

They have posted photos of this restaurant in their social media feeds.

You go there and find that the lighting is beautiful and the food is delicious.

Everything looks perfect, and the service is very attentive.

However, when you are sitting there eating, you might have a strange feeling.

You might feel that although the meal is tasty, it seems like you have eaten it somewhere before.

This feeling is very ordinary, without any special surprises.

Actually, this is not just your feeling alone.

This looks like a small matter because the AI is just helping you make a simple decision.

We feel this is convenient because we don't need to think for ourselves.

However, some scientists recently discovered that this might be a big problem.

There is a professor at Columbia University named Sandra Matz.

She and her team conducted a very important study.

She believes that when we let AI help us make decisions, we lose something.

We are not just handing over the task of finding a restaurant to AI.

We might also be handing over the important question of "who we are" to AI.

This professor calls this phenomenon the "becoming boring effect."

Her research found that when AI helps us decide what to buy, what to watch, and what to do, we change.

Although we become more efficient and do things faster, we also become more alike.

That is to say, the differences between everyone become smaller and smaller.

To prove this point, the professor and her students conducted a large survey.

They studied one thousand real social media users.

They analyzed eleven million choices made by these people.

This is a very large number, so the results are worth believing.

The researchers compared three different situations.

The first situation is where people rely completely on themselves to make decisions, without any computer help.

The second situation is using a regular AI assistant to help make decisions.

This regular AI doesn't understand your personal preferences; it just recommends based on general public data.

The third situation is using a personalized AI assistant.

This AI understands you very well; it knows what you liked and didn't like in the past.

The researchers were not looking at whether the things AI chose were good or not.

They mainly looked at the characteristics of people's choices.

They wanted to know how different your choices were from others' choices.

They also wanted to know if your own choices were always changing or always the same.

The results of the study were very clear, even a bit surprising to people.

Whether it is regular AI or personalized AI, they both have a common problem.

They both cause people to choose the most popular and ordinary things.

When we use AI, our choices become less special.

Regular AI makes everyone's choices become more or less the same.

Because it always recommends the things that are most popular according to statistical data.

For example, if everyone likes listening to pop music, AI will recommend pop music to you.

It thinks that what everyone likes must be the best.

So, the final result is that everyone's hobbies become the same.

Then, what about those personalized AIs that understand you?

Many people think that personalized AI should be better.

Because it knows what I like, and it will recommend things that suit me.

However, the study found that personalized AI also has a big drawback.

Although it recommends based on your habits, this makes you even less likely to change.

Your range of choices will become smaller and smaller.

For example, if you liked eating pasta before, it will keep recommending pasta to you.

It won't recommend Japanese or Chinese food because you haven't chosen them before.

In this way, you are unlikely to try new things yourself.

You will become very easy to predict because you are always doing repetitive things.

Simply put, regular AI makes us all become the same.

And personalized AI makes us become boring ourselves, no longer changing.

Both situations lead to one result, which is that human diversity decreases.

We lose that impulse to suddenly want to try new things.

We also lose the opportunity to discover unexpected surprises.

This impact is not just on the single matter of choosing a restaurant.

Now, AI is helping us make decisions in many aspects.

It decides what music we listen to on our phones.

It decides what movies we watch on video websites.

It even decides what news we see in news apps.

If AI always recommends the safest and most popular things to us, what will we become?

The professor believes this will affect our personalities and creativity.

Creativity often comes from those different, strange ideas.

If we only look at things everyone else looks at, it is hard for us to have new ideas.

We will be restricted to a very small circle.

This doesn't mean we should completely refuse to use AI.

AI is indeed a very useful tool, and it helps us solve many troubles.

However, the professor suggests that we need to design AI more smartly.

Future AI should not only pursue efficiency and maximum accuracy.

It should intentionally add a bit of "randomness" when making recommendations.

That is to say, it should occasionally recommend some things we haven't seen before.

Even if this thing might not have the highest rating, it is still worth a try.

Just like when we used to go to bookstores to buy books.

Sometimes you would buy a book just because the title sounded good, and it turned out to be very good.

This kind of accidental discovery is a very important part of life.

In the design of AI, besides safety and fairness, diversity should also be very important.

We must protect our individuality and not let everything become the same.

So, next time when your AI assistant recommends the "best" restaurant to you, you can stop and think for a moment.

You should ask yourself, what does "best" actually mean?

Many times, "best" is actually just "most ordinary" or "most visited by people."

This does not mean it is the most suitable for you, nor does it mean it is interesting.

True personalization should be helping you discover yourself, not limiting you.

In this digital age, efficiency is of course very important.

However, being a different person, being an interesting person, might be more important.

Perhaps we can try not listening to AI's advice occasionally.

We can make a decision ourselves, even if this decision looks a bit risky.

You can choose a restaurant with a lower rating or watch a movie no one knows about.

This might be a good way for you to rediscover your creativity.

We must remember that the right to choose is in our own hands.

Do not let algorithms completely decide your life.

Go choose those unexpected things, go walk those paths no one has walked.

Because only in this way can we maintain our characteristics as humans.

This is the topic about AI and choices that we wanted to chat about.

I hope you can recall this interesting study the next time you make a decision.

Keep your curiosity and try those different things more often.