斯坦福大学:我们需要人工智能的国家愿景(中英文)

译者注

谷歌前CEO 艾瑞克·施密特ERIC SCHMIDT 最近在纽约时报撰文,指出硅谷可能在AI等领域输给中国,他在文章开头就一针见血地指出,硅谷是美国创新的代名词,但硅谷的创新中有一个很重要的因素,那就是许多创新公司的领导者,包括谷歌的创始人,他们的起步都是从获得联邦政府的资助开始的。谷歌的创始人艾瑞克·施密特也是如此,我在1970年代和80年代在计算机科学领域的研究生工作,部分是由国家科学基金会(National Science Foundation)和美国国防高级研究计划局(Defense Advanced Research Projects Agency)资助的。

国家科学基金会近年来的财年预算都在七八十亿美金左右,NSF本身并不下设研究所或实验室,它以基金项目、合同和合作协议等形式,专门负责资助各个学科与工程领域中的基础研究和教育,约占联邦政府对基础研究总投入的1/4。许多发展中国家都在研究美国对于创新的研究机制和模式,但美国也在关注其他国家,艾瑞克·施密特的文章中蕴含着一种指向,或许美国的创新需要再次创新,所以他在文中提到应该建立一种合作伙伴关系,让大学研究人员和学生可以用上廉价的云计算。而斯坦福大学的新提议“国家研究云”(National Research Cloud)为该目标提供了一个思路。

斯坦福大学的提议是“We Need a National Vision for AI,标题很大,但文章却很简洁,在文章后半部的建议中,就首先指出,要鼓励协同,推动多元化的研究,宽带、互联网、人类基因组测序等重大创新都是基于政府和学术界的合作,未来的合作可能需要新的技术基础设施,比National Research Cloud ,这种基础设施可能为创新提供高质量的数据和高性能的算力,这种基础设施值得每年投入70亿美金。

 

 

 

我们需要人工智能的国家愿景

 

作者:李飞飞和约翰·埃特肯迪,2019

 

 

通过大胆的AI政策和计划建立全球领导地位对于我们的经济增长和社会稳定至关重要

如上,人工智能时代是黑暗而反乌托邦的——具有自我意识的杀手级机器人将会出现在我们所有人面前。这样的故事情节造就了轰动一时的电影,但他们对AI的发展速度以及存在的技术障碍缺乏真正的了解。

AI的时代正在到来,而且很快,在这其中有很多值得关注的事情,但终结者不是其中之一。

真正的威胁?包括美国在内的世界大多数国家都没有做好准备,以收获人工智能所提供的许多经济和社会利益,或减轻不可避免的风险。达到以上目标需要数十年。然而,在科学技术资金不足,支持力度不足甚至面临挑战的时代,人工智能应用的发展速度比我们的政策或机构要快。

这是一个全国性的紧急事件。

美国需要以人为中心的AI框架,以指导国家政策和计划——该框架以我们共同的美国平等,机会和代理价值观为基础:AI应该改善人类状况。  

我们需要大胆的领导,一个国家远景和一个以价值观为驱动力的国际标准,政策和原则框架。这需要国际,联邦,州和地方政府以及学术界,非营​​利组织和公司的空前合作和贡献。

华盛顿的早期努力——承诺到 2020年在多个机构和计划中投资近10亿美元用于研发;从2月开始生效的AI行政命令,该命令要求更新联邦机构的研究与开发策略,对数据开放的方法以及制定国际标准;以及在AI研究上进行适度投资的法案令人鼓舞,但还远远不够。

如果指导得当,人工智能时代将迎来一个人人享有生产力和繁荣的时代。普华永道(PWC)估计,到2030年,人工智能将为全球经济贡献15.7万亿美元。但是,如果我们不负责任地利用它并公平地分享收益,它将为那些迎接新时代的精英阶层带来更多的财富和权力集中-全球多数人的贫穷,无能为力和迷失了目标。

人工智能的潜在财务优势是如此之大,人工智能之间的鸿沟也是如此之深,以至于我们所知道的全球经济平衡可能会因一系列灾难性的构造转变而动摇。

随着赌注的增加,竞争将变得更加激烈。发达国家可能别无选择,只能推动自动化以提高生产率。麦肯锡认为,领先的人工智能国家可以额外获得20%-25%的净经济利益,而发展中国家则可能仅获得5%至15%。落后将对一个国家的前景造成灾难性的影响。即使是在足够有先见之明的地区可以投资人工智能的地区,由此产生的经济和阶级分化也可能导致危险的政治和社会冲突。

人工智能有能力成为我们最好(最坏)意图的力量倍增器。它可以帮助我们应对最棘手的挑战:管理自然资源;缓解气候变化;更早,更有效地发现和治疗疾病;照顾老人;为不断增长的全球人口增加粮食供应;物理和网络安全;高效的运输和基础设施;以及更多。但这也正在取代工作,甚至可能取代整个行业。加剧种族和性别在刑事司法,金融和就业方面的偏见;宣传和深造假货;以及对隐私和数据安全性的威胁越来越大。

我们已经看到,如果采用技术的时间快于管理技术的政策,并且没有事先考虑道德或负面影响,就会造成损害。最小化这些问题的最佳方法是让开发AI的人反映将受到AI影响的人群。

这个新框架必须优先考虑包容性和跨学科合作。 

我们的未来取决于社会和计算机科学家与来自不同背景的人们并肩工作的能力–与当今以计算机科学为中心的模型发生了重大转变。AI的创造者必须寻求跨种族、性别、文化和社会经济群体以及其他领域(例如经济学、法律、医学、哲学、历史、社会学、传播、人机交互、心理学和科学技术研究(STS)。这种协作应该贯穿于应用程序的整个生命周期,从最初的阶段到上市,直至其使用规模不断扩大。

为了执行以人为本的国家AI战略,我们建议美国政府建立一个涵盖教育,研究和企业家精神的新AI生态系统,在十年内至少投资1200亿美元。

具体来看:

1:支持公共研究,以追求下一代AI的突破,重点是跨学科研究。 

预算:70亿美元/年。

世界上许多最具影响力的技术–宽带、互联网、人类基因组测序——都是政府与学术界之间合作的结果。我们应该与领先的大学合作建立国家和地区研究中心,并强调跨学科研究和多元化的团队。

培训AI需要大量数据和强大的计算能力,这是大公司的主要优势,其激励结构促进了“点击”率而不是社会收益。公共组织之间的合作将确保为社会的利益而取得突破,启动“国家研究云”将为公共利益研究提供高价值的数据和高性能的计算。

2:投资教育,重点放在包容性上。 

预算:30亿美元(是目前联邦K-12 STEM年度支出的两倍)

PWC称,到 2030年代中期,将有多达30%的工作可以实现自动化,但是软件开发人员的就业预计在2016年至2026年之间将增长24%,其中不包括机器学习或机器人技术。2016年,只有不到15%的技术职位由黑人或西班牙裔人担任。女性仅占计算机角色的25%,仅占机器学习研究人员的12%。必须对此进行更改,以减轻偏见并构建受益于特权少数的应用程序。

美国需要对科学,技术,工程和数学(STEM)方面的未来劳动力进行教育,包括人工智能和计算机科学,并支持研究和解决就业转移和再就业的问题。

3:刺激创新并支持企业家。 

预算:$ 20亿美元

创业是我们经济的核心。小型企业与创业委员会估计,雇员少于100人的公司占美国企业的98%。他们推动创新和竞争,创造新的就业机会,并挑战垄断组织。我们应该通过捐赠、投资和技术资源为新兴技术提供早期支持,重点是农业、制造业、医疗保健、可持续性和清洁能源。

4:实施明确的、可操作的国际标准和准则,以指导人工智能的使用

与外国政府、公司和公民社会组织合作,以具体实施全球AI原则,例如OECD制定的原则。

在以人为中心的AI框架的开发方面引领世界对我们作为一个国家的理想至关重要,对于未来的经济增长和社会稳定至关重要。AI的技术发展必须与确保其负责任使用的政策共同发展。人工智能正在迅速发展,但如果我们现在就采取行动,我们仍有时间将它做好(纠正)。

FROM:https://hai.stanford.edu/news/we-need-national-vision-ai

本文由东西智库编译并校对,欢迎转载,转载须注明来源东西智库。

 

 

为避免翻译偏差,特附上英文版原文,供读者参考。

 

 

We Need a National Vision for AI

 

October 23, 2019

 

Establishing global leadership through a bold AI policy and plan is critical for the economic growth and stability of our society

According to the headlines, the Age of Artificial Intelligence is dark and dystopian — with self-aware, killer robots coming for us all. Such storylines make for blockbuster movies, but they lack a true understanding of AI, how quickly it’s developing and what technological barriers exist.

The Age of AI is coming, and fast, and there is plenty to be concerned about. But the Terminator isn’t one of them.

The real threat? Most of the world, including the United States, is unprepared to reap many of the economic and societal benefits offered by AI or mitigate the inevitable risks. Getting there will take decades. Yet AI applications are advancing faster than our policies or institutions at a time in which science and technology are being under-funded, undersupported and even challenged.

It’s a national emergency in the making.

The United States needs a human-centered AI framework that guides a national policy and plan — a framework anchored by our shared American values of equality, opportunity and agency: AI should improve the human condition.  

We need bold leadership, a national vision and a values-driven framework for international standards, policies and principles. This requires an unprecedented collaboration and commitment of international, federal, state and local governments, as well as academia, nonprofits and corporations.

Early efforts from Washington — a commitment to invest nearly $1 billion in R&D across multiple agencies and programs in 2020; the AI Executive Order from February, which called for an updated research and development strategy for federal agencies, an open approach to data and the development of international standards; and bills to make modest investments in AI research — are encouraging, but not nearly enough.

If guided properly, the Age of AI could usher in an era of productivity and prosperity for all. PWC estimates AI will deliver $15.7 trillion to the global economy by 2030. However, if we don’t harness it responsibly and share the gains equitably, it will lead to greater concentrations of wealth and power for the elite few who usher in this new age — and poverty, powerlessness and a lost sense of purpose for the global majority.

The potential financial advantages of AI are so great, and the chasm between AI haves and have-nots so deep, that the global economic balance as we know it could be rocked by a series of catastrophic tectonic shifts.

Competition will become fiercer as the stakes get higher. Developed countries might have no choice but to push automation to capture higher productivity. Leading AI countries could capture an additional 20%-25% in net economic benefits, while developing countries might see only 5% to 15%, according to McKinsey. Falling behind will have disastrous effects on a nation’s prospects. Even within regions prescient enough to invest in AI, the resulting economic and class divisions could lead to dangerous political and societal clashes.

AI has the ability to be a force multiplier of our very best — and very worst — intentions. It can help us address our most vexing challenges: managing natural resources; mitigating climate change; detecting and treating disease earlier and more effectively; caring for the elderly; increasing the food supply for a growing global population; physical and cybersecurity; efficient transportation and infrastructure; and much more. But it is also displacing jobs and possibly entire industries; exacerbating ethnic and gender bias in criminal justice, finance and employment; promoting propaganda and deep fakes; and increasing threats to privacy and data security.

We’ve seen the damage that can happen when technology is adopted faster than the policies that govern it, and without forethought about ethics or negative impacts. The best way to minimize these issues is for the people who develop AI to reflect the population that will be affected by it.

This new framework must prioritize inclusivity and interdisciplinary collaboration. 

Our future depends on the ability of social- and computer scientists to work side-by-side with people from multiple backgrounds —  a significant shift from today’s computer science-centric model. The creators of AI must seek the insights, experiences and concerns of people across ethnicities, genders, cultures and socio-economic groups, as well as those from other fields, such as economics, law, medicine, philosophy, history, sociology, communications, human-computer-interaction, psychology, and Science and Technology Studies (STS). This collaboration should run throughout an application’s lifecycle — from the earliest stages of inception through to market introduction and as its usage scales.

To execute a human-centered national AI strategy, we propose that the US government build a new AI ecosystem across education, research and entrepreneurship, with an investment of at least $120 billion over ten years.

Specifically:

1: Support public research to pursue the next generation of AI breakthroughs, with an emphasis on interdisciplinary research. 

Budget: $7 billion/ year.

Many of the world’s most impactful technologies — broadband; the internet; sequencing the human genome — have resulted from partnerships between government and academia. We should establish national and regional research hubs in partnership with leading universities and emphasize cross-disciplinary research and diverse teams.

Training AI requires vast stores of data and robust computing power, a major advantage to giant corporations whose incentive structures promote “click-through” rates over benefits to society. Collaboration between public organizations would ensure breakthroughs are developed in the interest of society and launching a National Research Cloud would provide high value data and high-performance computing for public-interest research.

2: Invest in education, with an emphasis on inclusion. 

Budget: $3 billion (double the current annual federal K-12 STEM spend)

As many as 30% of jobs could be automated by the mid 2030s, according to PWC, yet employment for software developers is expected to grow 24% between 2016 and 2026 — not including machine learning or robotics. In 2016, fewer than 15% of tech positions were held by black or hispanic people. Women comprise just 25% of computing roles and only 12% of machine learning researchers. This must change in order to mitigate bias and build applications that benefit more than a privileged few.

The US needs to educate a more diverse future workforce in science, technology, engineering and math (STEM), including artificial intelligence and computer science, as well as support research and programs to address job displacement and reskilling.

3: Spur innovation and support entrepreneurs. 

Budget: $2B

Entrepreneurship is the heart of our economy. The Small Business & Entrepreneurship Council estimates firms with fewer than 100 employees comprise 98% of US businesses. They drive innovation and competition, create new jobs and challenge monopolistic organizations. We should provide early-stage support for emerging technologies through grants, investment and technical resources, with an emphasis on agriculture, manufacturing, healthcare, sustainability and clean energy.

4: Implement clear, actionable international standards and guidelines for the ethical use of AI

Partner with foreign governments, companies, and civil society organizations to concretely implement global AI principles, such as those developed by the OECD.

Leading the world in the development of a human-centered AI framework is central to our ideals as a nation, and critical for the future economic growth and stability of our society. AI’s technical development must co-evolve with policies that ensure its responsible use. AI is advancing rapidly, but we still have time to get it right — if we act now.

FROM:https://hai.stanford.edu/news/we-need-national-vision-ai

 

注:本站文章除标明原创外,均未网友或机构投稿分享,如有宣发需求或侵权请联系dongxizhiku@163.com。