你會(huì)捐獻(xiàn)數(shù)據(jù)嗎？

There is a saying that scientists would rather share their toothbrushes than their data. But we will all have to imagine more open and hygienic forms of data sharing if we are to make the most of this revolution.

有句老話說(shuō)，科學(xué)家寧愿與人共用牙刷，也不愿意分享數(shù)據(jù)。但如果要充分利用這場(chǎng)革命，我們就必須設(shè)想更開(kāi)放和更衛(wèi)生的數(shù)據(jù)分享形式。

Data are emerging among the most important assets of the 21st century economy. Vast quantities of human- and machine-generated data, combined with ever more powerful computers and clever algorithms, are providing the feedstock for artificial intelligence systems and promising striking improvements in productivity. A country’s ability to exploit data in safe and creative ways will increasingly determine its success. It is high time for institutional innovation to encourage the process.

數(shù)據(jù)正成為21世紀(jì)經(jīng)濟(jì)最重要的資產(chǎn)之一。大量人工和機(jī)器生成的數(shù)據(jù)，加上越來(lái)越強(qiáng)大的電腦和聰明算法，正為人工智能系統(tǒng)提供原料并有望顯著提高生產(chǎn)率。一個(gè)國(guó)家以安全和創(chuàng)造性的方式利用數(shù)據(jù)的能力，將在越來(lái)越大程度上決定其成功。是時(shí)候讓制度創(chuàng)新鼓勵(lì)這一進(jìn)程。

For now, Pacific Rim companies, from the west coast of the US and the east coast of China, outstrip the rest of the tech industry. The likes of Google, Amazon, Alibaba, and Tencent are vacuuming up data and using them in inventive ways — to the delight of most consumers and the concern of some citizens worried about the concentration of corporate power and the erosion of privacy.

目前，環(huán)太平洋的公司(從美國(guó)西海岸到中國(guó)東部沿海地區(qū))超越科技行業(yè)其他公司。谷歌(Google)、亞馬遜(Amazon)、阿里巴巴(Alibaba)和騰訊(Tencent)正收集數(shù)據(jù)并以創(chuàng)新的方式利用數(shù)據(jù)，讓多數(shù)消費(fèi)者感到驚喜，也讓一些公民對(duì)企業(yè)權(quán)力集中和隱私被削弱感到擔(dān)憂。

Earlier this month, the UK government published an independent report by Professor Dame Wendy Hall and Jérôme Pesenti about how Britain could boost its own AI industry. Britain boasts an illustrious history in computing but risks becoming sub-scale in this emerging economy. The report made good suggestions for improving technical education, preserving the free movement of talented experts (a particular challenge after Brexit), and making government data readable by machines.

10月早些時(shí)候，英國(guó)政府發(fā)表了由溫迪•霍爾(Wendy Hall)教授和杰羅姆•佩森蒂(Jérôme Pesenti)撰寫的一份獨(dú)立報(bào)告，內(nèi)容是關(guān)于英國(guó)如何能促進(jìn)本國(guó)人工智能行業(yè)的發(fā)展。英國(guó)擁有輝煌的計(jì)算歷史，但有可能在新興的人工智能經(jīng)濟(jì)領(lǐng)域成不了氣候。報(bào)告提出了一些很好的建議，包括加強(qiáng)技術(shù)教育、保護(hù)專家人才的自由流動(dòng)(這在英國(guó)退歐之后尤其具有挑戰(zhàn)性)，以及讓政府?dāng)?shù)據(jù)變得機(jī)器可讀。

But the report’s most striking recommendation was to pioneer the creation of data trusts, a framework of agreements between government and industry to stimulate the secure and mutually beneficial exchange of data.

但報(bào)告最引人注目的建議是首創(chuàng)數(shù)據(jù)信托：政府與行業(yè)之間的一個(gè)協(xié)議框架，旨在推動(dòng)安全和互利的數(shù)據(jù)交換。

It is a neat an idea in theory but it will be hard to implement in practice. Even Professor Hall admits: “There is a will to do it but not yet a way.”

這在理論上是一個(gè)高明的想法，但在實(shí)踐中將很難落實(shí)。就連霍爾教授也承認(rèn)：“目前有這么做的意愿，但還沒(méi)有找到方法。”

The central idea would be for the government to work with independent institutions, such as the Royal Society and the Open Data Institute, to encourage the creation of robust arms-length mechanisms to share public and private sector data. That would help reassure data donors that their information was being used for the collective good rather than for private profit or government surveillance. But it would also give researchers and entrepreneurs access to exciting fresh seams of data.

中心設(shè)想是政府與獨(dú)立機(jī)構(gòu)合作，例如英國(guó)皇家學(xué)會(huì)(Royal Society)和開(kāi)放數(shù)據(jù)研究所(Open Data Institute)，以鼓勵(lì)創(chuàng)建穩(wěn)健的半官方機(jī)制，分享公共和私營(yíng)部門數(shù)據(jù)。這將有助于讓數(shù)據(jù)捐獻(xiàn)者放心：他們的信息將被用于集體利益，而不是用于牟取私利或政府監(jiān)視。但這也將讓研究人員和企業(yè)家獲得令人興奮的新數(shù)據(jù)源。

The report’s authors argue the public sector could benefit enormously in realms such as energy, healthcare, transport, and cyber security. Other experts suggest that regulators could incentivise companies, like Uber, to share anonymised data, helping urban planners, for example. Charitable research institutions, such as the Wellcome Trust, could encourage pharmaceutical companies to contribute “exhaust” data from failed clinical trials that could provide invaluable insights.

報(bào)告的作者們辯稱，公共部門有望在能源、醫(yī)療、交通和網(wǎng)絡(luò)安全等領(lǐng)域大大受益。其他專家提出，監(jiān)管者可以鼓勵(lì)優(yōu)步(Uber)等公司分享經(jīng)過(guò)匿名處理的數(shù)據(jù)，比如幫助城市規(guī)劃者。維康信托(Wellcome Trust)等慈善研究機(jī)構(gòu)可以鼓勵(lì)制藥公司貢獻(xiàn)來(lái)自失敗臨床試驗(yàn)的“排出”數(shù)據(jù)，它們可能提供非常寶貴的洞見(jiàn)。

Sir Nigel Shadbolt, AI professor and co-founder of the ODI, says the aim should be to create “interested, disinterested institutions” to supervise such communal data assets. He cites the historic example of the Lloyd’s Register of Shipping, a charitable organisation dating back to the 1760s, set up to accredit and classify merchant ships for the benefit of all trading companies and insurers. Copyright libraries in the UK and Ireland have also been granted the right to obtain all published books for free to spread knowledge.

人工智能教授、英國(guó)開(kāi)放數(shù)據(jù)研究所聯(lián)合創(chuàng)始人奈杰爾•沙德博爾特爵士(Sir Nigel Shadbolt)表示，目標(biāo)應(yīng)該是創(chuàng)建“感興趣且排除自身利益的機(jī)構(gòu)”，來(lái)監(jiān)管這些共享數(shù)據(jù)資產(chǎn)。他引用了可追溯到18世紀(jì)60年代的勞氏船級(jí)社(Lloyd’s Register of Shipping)這個(gè)具有歷史意義的例子，這個(gè)慈善組織創(chuàng)建的目的是為所有貿(mào)易公司和保險(xiǎn)公司服務(wù)，對(duì)商船進(jìn)行認(rèn)證和分類。英國(guó)和愛(ài)爾蘭的版權(quán)圖書(shū)館也被授予免費(fèi)獲得所有已出版書(shū)籍的權(quán)利，以傳播知識(shí)。

“We have been here before and have quite successfully established regimes where information has been made available for the common good in perpetuity,” he says. “The idea of creating data trusts is well worth pursuing.”

“我們以前曾經(jīng)這么做過(guò)，并成功建立了制度，為了共同利益讓信息永久可獲取，”他表示，“創(chuàng)建數(shù)據(jù)信托的想法是非常值得探索的。”

Privacy and security are the most obvious obstacles. Anonymising data is more difficult than it sounds, particularly for health information. The EU’s sweeping General Data Protection Regulation, which comes into force in May and will be adopted by Britain, also imposes strict restrictions on data use. “GDPR is fantastically protective of people but potentially restrictive for research,” says Professor Hall.

隱私和安全是最明顯的障礙。數(shù)據(jù)匿名化要比聽(tīng)上去更難，尤其是健康信息。歐盟覆蓋面極廣的《一般數(shù)據(jù)保護(hù)條例》(General Data Protection Regulation)——今年5月生效，英國(guó)將施行該條例——也對(duì)數(shù)據(jù)使用設(shè)置了嚴(yán)格限制?；魻柦淌诒硎荆?ldquo;條例很好地保護(hù)了人們，但潛在可能限制研究。”

That may well put Europe at a further competitive disadvantage in the short term given looser accountability elsewhere. Companies in China, in particular, face few constraints on data use, either from government regulation or public concern.

鑒于其他司法管轄區(qū)的問(wèn)責(zé)制度較為寬松，這很有可能會(huì)讓歐洲在短期內(nèi)處于進(jìn)一步的競(jìng)爭(zhēng)劣勢(shì)。尤其是，中國(guó)企業(yè)在數(shù)據(jù)使用方面幾乎沒(méi)有受到任何制約，包括來(lái)自政府法規(guī)和公眾擔(dān)憂的制約。

But Jürgen Schmidhuber, professor of artificial intelligence at the University of Lugano, argues that safeguarding data rights may make good business sense over the longer term if it results in a more trusted infrastructure for data sharing. “Respect for privacy will be profitable one day,” he says.

但盧加諾大學(xué)(University of Lugano)人工智能教授于爾根•施密德胡伯(Jürgen Schmidhuber)辯稱，如果帶來(lái)更受信任的數(shù)據(jù)分享基礎(chǔ)設(shè)施，那么保護(hù)數(shù)據(jù)權(quán)利可能會(huì)在較長(zhǎng)期有利于商業(yè)。他表示：“終有一天，對(duì)隱私的尊重將帶來(lái)盈利。”

[email protected] 譯者/梁艷裳