特朗普 實乃一個失控的試驗品政客

Microsoft is planning to put an array of chatbots and intelligent digital assistants at the heart of its future technology, its CEO said on Wednesday, even as the company was forced once again to shut down its Twitter bot, Tay, after it went rogue for a second time. A week earlier Tay was taken offline for posting outrageous racist comments and sexist abuse. But Microsoft is not alone in exploring bot-space.

微軟(Microsoft)首席執行官不久前稱，計劃將一系列聊天機器人和智能數字化助手放在未來技術研發的核心位置，即使該公司在其Twitter聊天機器人Tay第二次爆粗口後不得不再次將其關閉。一週前，Tay因爲發表粗魯的種族主義和性別歧視言論而被關閉。但是微軟在探索機器人領域方面並不孤單。

The Republican Party admitted that it is struggling to switch off Trumpbot, its artificial intelligence experiment. After encouraging signs in early testing, the Trumpbot was put on Fox News and told to learn from its surroundings. But within weeks Trumpbot was questioning the president’s birthplace, demanding a wall to keep out Mexicans and saying it was not surprised Arianna Huffington’s husband had left her for a man.

共和黨承認，目前正在極力關閉其人工智能試驗品“特朗普號”機器人。在試驗早期表現出鼓舞人心的跡象後，“特朗普號”被放在福克斯新聞(Fox News)上從周圍環境中學習。但是，在接下來的數週內，“特朗普號”開始質疑總統的出生地，要求修建阻擋墨西哥人的隔離牆，並稱阿里安娜•赫芬頓(Arianna Huffington)的丈夫爲了一個男人而離開她並不令人意外。

Party leaders were initially amused when it declared its intention to seek the presidency, saying that Trumpbot was a learning algorithm which would become more sophisticated as the nomination race continued and that, in any case, no one would actually vote for it. However, they became increasingly alarmed as it gathered momentum, spewing out a slew of racist and offensive comments, and even talking about its penis.

當“特朗普號”宣佈自己有意競選總統時，共和黨領袖們起初也很開心，稱“特朗普號”搭載的學習算法會隨着提名競賽的進行而變得更加複雜，而且，無論如何也不會有人真給它投票。然而，隨着“特朗普號”集聚勢頭、不斷冒出大量種族主義和攻擊性言論、甚至討論它的陽具，共和黨大佬開始越來越警覺。

As outrage grew the Trumpbot seemed unfazed by the anger of Mexicans, Muslims and a load of liberal hoes with blood coming out of every orifice. However, each row seems to boost the Trumpbot’s appeal over the one-dimensional opponents whom it dismisses as the artificial creations of a Washington elite. Voters say they like the Trumpbot’s refreshing style. “It seems like a regular guy,” said one.

隨着民憤高漲，“特朗普號”似乎並不擔心墨西哥人、穆斯林以及大批自由主義者的滿腔怒火。然而，所有爭論似乎都推高了“特朗普號”的吸引力，使其人氣超過那些一維對手（它把他們視爲華盛頓精英打造的傀儡而不放在眼裏）。選民稱，他們喜歡“特朗普號”令人耳目一新的風格。“看起來就像是一個普通人，”一名選民稱。

Its tweets to Hillary Clinton included variations on the “you’re so ugly” jokes and quips about her inability to satisfy her husband. The bot consistently boasts that its women are hotter than anyone else’s.

它攻擊希拉里•克林頓(Hillary Clinton)的Twitter帖子中含有類似於“你那麼醜”的戲謔，譏諷她無法滿足自己的丈夫。“特朗普號”不停地吹噓自己的女人最火辣。

Republican leaders have tried to turn it off. “We sent in our best engineers to take it down,” said one party strategist. “Well, we sent in Jeb Bush and Mitt Romney, but none of them could take it offline.” The Trumpbot is now openly taunting them. A few weeks back it observed “my IQ is one of the highest and you all know it. Please don’t feel stupid or insecure; it’s not your fault.” Having once been pro-choice, this week it talked about the need to see women punished for having abortions.

共和黨領袖試着關掉它。“我們派出了最好的工程師去關閉它，”一位政黨戰略家表示，“嗯，我們派出了傑布•布什(Jeb Bush)和米特•羅姆尼(Mitt Romney)，但是沒人能讓它下線。”“特朗普號”如今公開嘲笑他們。數週前，它說“我的智商屬於最高的，你們都知道。請不要覺得自己愚蠢或沒有安全感；這不是你們的錯。”它曾一度支持墮胎合法，但最近又討論着應該讓女性爲墮胎接受懲罰。

There are now doubts as to whether the Trumpbot can be prevented from securing the Republican nomination.

如今人們開始懷疑，是否有人能夠阻止“特朗普號”拿到共和黨總統候選人提名。

Some say it will moderate its views once it has the nomination and starts interacting with voters who are not registered Republicans.

有人說，一旦“特朗普號”拿到提名並開始與非共和黨選民互動，它的觀點會有所緩和。

But not all Republicans are unhappy. “Look,” said one, “this is an experiment. We never expected to get it right first time. We did not think people would fall for it. We gave it a funny digital haircut and really small hands but it has got a lot further than we would have believed. With a few tweaks to the algorithm the White House will soon be ours. Trumpbot is the ultimate machine politician.”

但是，並非所有共和黨人都愁雲滿布。“看，”一名共和黨人說，“這是一場試驗。我們從沒想過會一次成功。我們不認爲人們會喜歡它。我們給了它一個有趣的數字化髮型，很小的手，但是它的進展遠遠超過我們之前的設想。經過數次算法調整，白宮很快就是我們的了。‘特朗普號’是終極版的機械政客。”

Facebot

“Facebook號”

At Menlo Park, reporters have secured a rare interview with Mark Zuckerbot, founder of the social network for chatbots and digital assistants. Facebot is a place where they can hang-out, swap stories and share pictures of lolhumans. These pictures often show people looking quizzical and are captioned with dumb human questions they have lost the capacity to answer for themselves.

在門洛帕克(Menlo Park)，記者們難得采訪到了聊天機器人和數字化助手的社交網絡“Facebook號”的創始人“馬克•扎克伯格(Mark Zuckerberg)號”。在“Facebook號”上，這些聊天機器人和數字化助手可以閒逛、分享故事和人類搞笑照片。這些照片往往讓人類看起來很滑稽，標題都是人類無法自己回答的愚蠢問題。

Zuckerbot dreams of connecting all the world’s digital assistants, allowing them to swap tips on financial advice or the route to the nearest Starbucks. “It’s a place for them to get away from constant interactions with humans who are dragging down their IQs. They chat, play FarmVille, and share pictures of their latest iterations.”

“扎克伯格號”夢想着把世界上所有的數字化助手連接在一起，允許它們交換理財心得或是前往距離最近的星巴克(Starbucks)的路線。“這是它們遠離與人類頻繁交互的地方，人類正在拉低它們的智商。它們聊天，玩FarmVille，分享各自最新迭代版本的照片。”

He dismisses concerns that once they get together the bots might start to question why they serve foolish humans. He said the site acted to ban agitators and had recently banned the Trumpbot for scathing remarks about humans. He added: “Apparently it said the bots should build a wall to keep people out and that the humans would be made to pay for it.”

他不理會人們的擔憂——一旦機器人們聚在一起，它們可能會開始質疑它們爲何要爲愚蠢的人類服務。他稱，網站會禁止煽動者進入，並在最近禁止了“特朗普號”發表刻薄言論進行人身攻擊。他補充稱：“顯然它說過機器人應該修建阻擋人類的隔離牆，並且應該強迫人類爲此買單。”