Name: 哈佛大學人工智能獵殺外星人項目內幕 (Inside the AI Alien Hunting Project at Harvard)
Uploaded: 2025-08-08T04:56:09.000Z
Duration: 15 min 14 s

Grainy military footage showing objects defying explanation.

模糊的軍事鏡頭顯示了一些無法解釋的物體。

In 2017, Pentagon officials confirmed the videos were real.

2017 年，五角大樓官員證實這些視頻是真實的。

I think it is arrogant of us to think that we are alone in the universe.

我認為，我們認為自己在宇宙中是孤獨的，這是狂妄自大。

While the government has kept its investigation mostly classified, a small team of scientists at Harvard launched their own, building artificial intelligence to scan the sky for UAP and sharing everything they find.

雖然政府將其調查大部分列為機密，但哈佛大學的一個科學家小組卻發起了他們自己的調查，建立人工智能來掃描天空中的 UAP，並分享他們發現的一切。

Humanity came to exist only over the past few million years.

人類是在過去幾百萬年中才出現的。

That's one part in 10,000 of the age of the universe.

這相當於宇宙年齡的萬分之一。

And we know that we are not at the center of the universe.

我們知道，我們不是宇宙的中心。

So if you arrive to a play, and you arrive just at the end of the play, and you are not at the center of stage, the play is not about you.

是以，如果你去看一齣戲，而你恰好在戲的結尾到達，並且你不在舞臺的中心，那麼這齣戲就與你無關。

And it might be the most serious scientific effort yet to answer the question, are we alone?

這可能是迄今為止回答 "我們是否孤獨 "這一問題的最嚴肅的科學努力。

In 2021, the U.S. government released an official report confirming more than 140 sightings of unidentified aerial phenomena.

2021 年，美國政府發佈了一份官方報告，確認了 140 多起不明空中現象目擊事件。

Shortly after, Avi Loeb launched the Galileo Project, a scientific effort to investigate what the government could not.

不久之後，阿維-勒布發起了伽利略計劃，通過科學努力調查政府無法調查的事情。

Two years ago, we started looking into this story.

兩年前，我們開始調查這個故事。

We had heard that Harvard was embarking on a project to build artificial intelligence software and deploy this observatory housed at an undisclosed location in somewhat rural Massachusetts to begin trying to detect UAP or UFOs.

我們聽說哈佛大學正在開展一個項目，建立人工智能軟件，並在馬薩諸塞州一個未公開的地點部署這個觀測站，開始嘗試探測UAP或UFO。

I'm a professor of science at Harvard University, the director of the Institute for Theory and Computation, and the head of the Galileo Project at Harvard University.

我是哈佛大學理學教授、理論與計算研究所所長，也是哈佛大學伽利略項目的負責人。

The Galileo Project is a series of scientific projects aimed at exploring the possibility of extraterrestrial life.

伽利略計劃是一系列旨在探索外星生命可能性的科學項目。

There were a few aspects of this research that really interested me as a technology and cybersecurity reporter.

作為一名技術和網絡安全記者，我對這項研究的幾個方面非常感興趣。

This topic of UAP, you know, fits very neatly into this category of national security threat.

要知道，UAP 這一話題完全符合國家安全威脅的範疇。

You also have the AI and technology issue to potentially detect and potentially mitigate a very serious national security threat.

此外，人工智能和技術問題也有可能偵測並減輕非常嚴重的國家安全威脅。

We felt like this story and this research, you know, really dived into an interesting application of AI software, the likes of which we hadn't seen before.

我們覺得這個故事和這項研究，你知道，真正深入到了人工智能軟件的有趣應用中，這種應用我們以前從未見過。

So, Professor Loeb, do you want to tell us a little bit about where we are and what we're looking at?

那麼，勒佈教授，你想告訴我們一點關於我們的現狀和我們正在研究的問題嗎？

我們來到了第一個伽利略天文臺。

We are testing all the instruments, and then we will make a copy of it and place it in many locations.

我們正在對所有儀器進行測試，然後會複製一份，放在許多地方。

A network of cameras and specialized sensors is recording the entire sky 24 hours a day.

由攝影機和專用傳感器組成的網絡每天 24 小時記錄著整個天空。

That data is then fed into artificial intelligence software, built from scratch, and trained to filter out planes, birds, and other known objects.

然後，這些數據會被輸入從零開始構建的人工智能軟件，並經過訓練以過濾掉飛機、鳥類和其他已知物體。

The goal is to identify any anomalies and collect as much data as possible about them.

目標是找出任何異常情況，並收集儘可能多的相關數據。

When I think of an observatory, I think of a giant telescope on the top of a hill.

一提到天文臺，我就會想到山頂上的巨型望遠鏡。

I don't think of a patch of blacktop with some instruments that I've never seen before.

我不認為一塊黑土地上有一些我從未見過的儀器。

It might look unusual to people who are seeing it.

這可能會讓看到它的人覺得不尋常。

If you have just one giant telescope, then you're looking at a small portion of the sky, very often smaller than the size of the moon.

如果你只有一個巨型望遠鏡，那麼你看到的只是天空的一小部分，往往比月球還小。

As a result, you might miss objects crossing the sky.

是以，您可能會錯過劃過天空的物體。

And that's what happens in astronomical observations because they don't care about things near Earth.

這就是天文觀測的結果，因為他們並不關心地球附近的事物。

它們關注的是非常遙遠的信號源。

But we care about objects passing overhead, and for that, we need to look at the entire sky at all times.

但我們關心的是從頭頂掠過的物體，為此，我們需要隨時觀察整個天空。

A group of Harvard researchers is leading the development of the AI software powering the Galileo project.

哈佛大學的一組研究人員正在牽頭開發為伽利略項目提供動力的人工智能軟件。

I was interested in the topic of UAPs since I was in high school.

我從高中開始就對 "UAPs "這一主題很感興趣。

UAP stands for Unidentified Anomalous Phenomena.

I was very intrigued by all the stories and the patterns that seemed to be coming up again and again.

我對所有的故事和似乎一再出現的模式非常感興趣。

And it was a little bit of a mystery to me why that wasn't a topic worthy of research.

我有點不明白，為什麼這不是一個值得研究的課題。

Ultimately, as scientists, you should be curious about the world around you.

歸根結底，作為科學家，你們應該對周圍的世界充滿好奇。

And data is what answers the questions that you come up with.

而數據就是你提出問題的答案。

Richard Chloetzi and Laura Domine are postdoctoral researchers at Harvard.

Richard Chloetzi 和 Laura Domine 是哈佛大學的博士後研究員。

They are full-time staff assigned to the Galileo project.

他們是伽利略項目的全職工作人員。

They are both accomplished scientists in their own right.

These are the kind of scientists who have the world at their fingertips in terms of potential employment opportunities.

這些科學家擁有全世界的潛在就業機會。

What's really interesting about this story to us is that they've chosen to lean into their personal interests in the topic of UAP and the search for extraterrestrial life.

對我們來說，這個故事真正有趣的地方在於，他們選擇了在 UAP 和尋找外星生命的話題上傾注自己的個人興趣。

They want to make this work of attempting to detect UAP something that's more accepted mainstream.

他們想讓這項試圖檢測 UAP 的工作成為更被接受的主流。

We do have sensors pointed at the sky in different locations.

我們確實在不同地點安裝了對準天空的傳感器。

The military is looking at it from satellites and from the ground.

軍方正在從衛星和地面進行觀測。

The big difference is they're always looking for something that is known.

最大的區別在於，他們總是在尋找已知的東西。

And we are looking to classify everything that is known so that we can focus on the unknowns.

我們希望對已知的一切進行分類，這樣就能集中精力解決未知問題。

For building models that are able to detect UAP, it's quite a long, involved process.

對於建立能夠檢測 UAP 的模型來說，這是一個漫長而複雜的過程。

First of all, you need a model that is trained and capable of detecting objects in the sky, classifying them.

首先，你需要一個經過訓練的模型，它能夠探測天空中的物體並對其進行分類。

And even just to get into that step has been a monumental task because there's currently no models out there which are capable of detecting objects in the sky with high accuracy.

即使是進入這一步也是一項艱鉅的任務，因為目前還沒有任何模型能夠高精度地探測天空中的物體。

Most models have been built for detecting and classifying things on the ground.

大多數模型都是用於探測地面上的物體並對其進行分類。

And so we had to really go ahead and train a model from scratch.

是以，我們不得不從頭開始訓練一個模型。

What we are seeing here is a pan-tilt-zoom camera, basically a security camera that we are developing software for so that it will zoom on objects of interest in the sky.

我們在這裡看到的是一臺平移-傾斜-變焦攝影機，基本上是一臺我們正在開發軟件的安全攝影機，這樣它就能變焦拍攝天空中感興趣的物體。

This is a spectrometer attempting to detect any emission from sources that transmit radio waves.

這是一個光譜儀，試圖探測無線電波發射源的任何發射。

We're building an array of instrumentation that we are deploying over a pretty large area of land.

我們正在建造一個儀器陣列，部署在相當大的陸地範圍內。

And these instrumentations monitor the entire electromagnetic spectrum.

這些儀器可以監測整個電磁頻譜。

We're also looking at a different distance.

Usually observatories look at space, right?

And we are focused on objects within the atmosphere, so in low altitude.

而我們關注的是大氣層內的物體，所以是在低空。

The main workhorse is actually this set of infrared cameras that we call the DALE, a collection of eight cameras looking at the entire sky.

主要的工作設備實際上是這套我們稱之為 DALE 的紅外攝像機，它由八臺相機組成，可以觀測整個天空。

Everything we're doing here is brand new, so to speak.

可以說，我們在這裡所做的一切都是全新的。

And even if at the end of it we find nothing, it's still a result.

即使最後一無所獲，也是一個結果。

We're still finding that there's nothing there.

我們仍然發現那裡什麼都沒有。

So obviously we prefer to find something anomalous, but also just the whole process of actually developing these technologies and measuring the entire electromagnetic spectrum is an incredibly important task in and of itself.

是以，我們顯然更希望發現一些異常現象，但實際開發這些技術和測量整個電磁頻譜的整個過程本身就是一項極其重要的任務。

To date, the team has tracked and classified hundreds of thousands of objects, most of them explainable.

迄今為止，研究小組已經追蹤並分類了數十萬個物體，其中大部分都是可以解釋的。

The system is still new, and the search, they say, is just getting started.

他們說，這個系統還很新，搜索工作才剛剛開始。

But Avi has become a polarizing figure because of this work.

但阿維卻因為這項工作而成為一個兩極分化的人物。

His interest in the topic began in 2017, after a telescope in Hawaii spotted a strange, football-field-sized object in space called Oumuamua.

他對這一話題的興趣始於 2017 年，當時夏威夷的一架望遠鏡在太空中發現了一個足球場大小的奇怪物體，名為 Oumuamua。

A lot of academia works on reputation, and there has been a lot of stigma on this topic that people are not ready to shake off.

學術界的很多工作都是為了聲譽，在這個問題上，人們還沒有準備好擺脫很多汙名。

People are entitled to criticize all they want.

Ultimately, data is what's going to solve this mystery.

最終，數據才能解開這個謎團。

I get attacked personally on the research I'm doing, and it clearly touches a nerve in our human existence to admit that we might not be the smartest in the universe, that there might be someone else out there.

我所做的研究受到了人身攻擊，承認我們可能不是宇宙中最聰明的，可能還有其他人存在，這顯然觸動了我們人類生存的神經。

To me, the highest level is direct evidence.

對我來說，最高級別的證據是直接證據。

It has been what I was hoping to do, which is to take sensors, calibrate them, and put them out there in the field in service of UAP research.

這正是我所希望做的，即使用傳感器，校準它們，並將它們投入實地，為 UAP 研究服務。

And in the process, showcasing to our colleagues in academic research that this is something you can do while following the scientific method.

在這一過程中，向我們的學術研究同行展示，在遵循科學方法的同時，這也是可以做到的。

The point of this section is to show how we would rigorously quantify how many objects we count in our dataset.

本節的重點是說明我們如何嚴格量化數據集中的對象數量。

So in this example, over five months, we collected about half a million objects that we reconstructed with a 95% confidence level.

是以，在這個例子中，在五個月的時間裡，我們收集了大約 50 萬個對象，並以 95% 的置信度進行了重建。

So this is an example of rigorous analysis that we plan to repeat for other instruments.

是以，這是一個嚴謹分析的例子，我們計劃對其他儀器進行重複分析。

The stigma attached to UAP is real and powerful and challenges national security.

UAP 所帶來的恥辱是真實而強大的，對國家安全構成了挑戰。

In recent years, there has been increasing public disclosure of government involvement in the unidentified anomalous phenomena topic.

近年來，越來越多的公眾披露了政府參與不明異常現象專題的情況。

Numerous pilots and government officials have come forward to describe their experiences, and Congress has elevated the issue to a top national security concern.

許多飛行員和政府官員都站出來講述了他們的經歷，美國國會也將這一問題提升為國家安全的首要問題。

As we convene here, UAP are in our airspace, but they are grossly underreported.

當我們在這裡召開會議時，UAP 就在我們的領空，但對它們的報告卻嚴重不足。

字幕列表影片播放

哈佛大學人工智能獵殺外星人項目內幕 (Inside the AI Alien Hunting Project at Harvard)

perspective

ultimately

entire

incredibly