打开

李开复《时代》百大人物峰会震撼演讲:人工智能就是新时代的“电力”(附视频&演讲稿)

精彩英语演讲 2019-04-26 10:06

点击上方“精彩英语演讲”,选择“设为星标”

英语演讲视频,第一时间观看

4月25日,创新工场创始人李开复受邀参加百大人物峰会并发表《Making A.I. Work for Humans(让AI造福人类)》主题演讲,探讨当今人工智能发展现状,分享了人工智能所带来的机遇与挑战。

在《让AI造福人类》的演讲中,李开复谈到,人工智能最大的突破是深度神经网络,也称为深度学习。深度学习通过某个单一领域和大量的训练数据进行工作,没有人为干预自行进行决策、预测或分类,这项技术需要的是大量的数据。

目前深度学习已经广泛以用在游戏、自然语言理解、标准化测试、医疗保健应用中的放射检查等,在以上这些领域中,人工智能做得比人类更出色。

同时,李开复把人工智能看作新时代的“电力”,他认为:“人工智能像电力一样,会赋能驱动很多应用和设备的高速发展,其中很多会超越人类想象。也正是因为它是像电力一样的存在,多年以后,我们也会无法想象没有人工智能的生活会是怎样。几乎所有能想象到的领域都会被影响。”

李开复《时代》周刊百大人物峰会演讲

上下滑动,查看英文演讲稿

图片文字来自:量子位

打开网易新闻 查看更多图片

AI expert and Sinovation Ventures CEO Dr. Kai-Fu Lee took the stage at the TIME 100 Summit in New York on Tuesday to discuss the current state of artificial intelligence, arguing that it will do as much for society’s well-being as electricity. In his talk, Lee also touched on how governments and businesses can take advantage of artificial intelligence to improve nearly everything from infrastructure to the relationship between individuals and the companies handling their data.

Lee showcased examples of AI applications, like warehouse machines that use computer vision to move and sort boxes, and an AI-powered rapper capable of generating rhymes based on any topic. Lee also highlighted an AI technique known as deep learning, and how it could be and has been used to generate false information, demonstrating with a fabricated audio clip of President Donald Trump. New machine learning techniques, like generative adversarial networks, have created real images of nonexistent people, as well as legit-looking news stories full of false information.

Artificial intelligence generating fabricated audio, images, or stories is a major concern, especially when combined with the ongoing rise of fake news. “What’s more, biases inherent in machine learning software will continue to perpetuate an inequity between minorities and underrepresented communities,” said Lee, a comment addressing lack of training data for underrepresented communities, leading to a marked decrease in accuracy and effectiveness for those users.

He also showed how AI is capable of providing a positive impact in China, where teachers in impoverished areas of the country are able to take advantage of the technology to grade tests and assign homework. “It’s an experience that will break down the barrier to education for those who need it,” said Lee. (Of course, AI is also being used in China and elsewhere for facial recognition security software that some are concerned can violate citizens’ rights.)

Privacy, a hot-button issue in a time when user data gathered by major corporations is mismanaged or abused, is also on the list of problems artificial intelligence can tackle, Lee argues. “Certainly regulations are needed to prevent the most egregious misuse of data,” said Lee. But at the same time, Lee thinks technologists will eventually develop AI-powered tools to protect users while giving them the freedom to choose their level of privacy. “We can imagine a slider that gives each individual a choice, that gives an individual more security or more convenience,” said Lee, equating the idea to companies in the past that once created anti-virus software for computers.

He also discussed job displacement occurring alongside the proliferation of artificial intelligence. “I think governments need to think about education, training, and changing the society’s belief about compassion and work,” said Lee, who thinks redesigned vocational programs and a focus on “compassionate work” should be considered when it comes to training a workforce.

As for jobs, Lee isn’t shy about the coming disruption thanks to artificial intelligence. Customer service, telemarketing, and accounting jobs will all be subject to “serious job displacement issues,” according to Lee, and finding new careers for workers in the field will require difficult retraining. But Lee thinks, like electricity, AI’s potential to revolutionize every aspect of our daily life is more tangible than theoretical. “AI is here to stay, and i think we need to rise up to the occasion and embrace it.”

演讲中,李开复表示,与电力不同,人工智能游四个大型的加速器,能比电力更快获得运用。首先,世界上有两大人工智能超级大国,美国和中国正在共同推动人工智能向前发展;其次强大的互联网巨头利用大量数据培训了数万名AI工程师;第三有许多人工智能基金,包括创新工场、软银基金等正在推动人工智能向前发展;最后,AI正在成为一个平台,人工智能的编程正在变得越来越容易,而且在接下来的15年里它会变得更为日常,任何工程师都可以在云上编程AI。这与需要花数十年建立电网的电力相比大为不同。

随着时间的推移,人工智能将从一项复杂的科学演变成为一项B2B技术,并逐渐成为每个工程师都可以使用的主流技术,每个企业都会拥抱和利用它。公司将基于此节约成本、提高效率、使用新的应用程序,从而实现巨大的经济利益。

演讲中,李开复通过一些实例展示了当下人工智能技术的发展现状,包括基于计算机视觉来移动、分拣物体的仓储机器人,以及能够根据任何主题进行创作和表演的AI说唱歌手、教神经网络如何像某个人一样说话等等。

李开复肯定了人工智能技术在医疗健康、教育等领域的积极作用。如在中国最贫困的地区,利用人工智能技术加上直播技术的辅助,可以把农村学校与身处北京的顶尖教师们连接起来,让这些孩子们享受到优秀的教师资源。

另外,李开复博士也谈到了人工智能带来的隐私安全、就业机会等一系列问题。

他表示,为了防止最严重的数据滥用,需要制定相应的法规。但是,人们不应该只将人工智能带来的隐私问题视为一个监管问题,可尝试用“以子之矛攻己之盾”——用更好的技术解决技术带来的挑战,例如同态加密、联邦学习等技术。

对于伴随人工智能普及而发生的人类工作被取代问题,他也发表了自己的见解——政府需要切实关注教育和培训,人们也需要在工作岗位中发挥自己的创造力和同情心——这是人们真正擅长的,也是机器人无法取代的。

往期精彩英语演讲集

外媒全英文专访李开复和李飞飞:人工智能时代,人性才是最热门的技能(附双语视频)
李开复TED流利英文演讲:AI时代失去工作不可怕,可怕的是失去生活的意义(附视频&演讲稿)
李开复剑桥大学全英文演讲:人工智能的市场价值将超过整个互联网世界(附视频)
李开复哥伦比亚大学全英文演讲:爱,让人类有别于人工智能(附视频&演讲稿)
Bloomberg专访李开复:硅谷创业者根本不是中国创业者的对手(附视频)
李开复达沃斯全英文专访:ICO才是2018年最大的泡沫(附视频&演讲稿)

防止未来失联


特别声明:本文为网易自媒体平台“网易号”作者上传并发布,仅代表该作者观点。网易仅提供信息发布平台。
帮TA点赞
大家都在看打开应用 查看全部
网易热搜每30分钟更新
打开应用 查看全部
打开