人工智能应该被政府管制么?

tinyfool 发布于 2017年09月16日 | 更新于 2017年09月16日
无人欣赏。

作者:Andrew Ng

译自:quora

AI作为基本技术不应受到管制。政府可能也没办法阻止你在笔记本电脑上玩神经网络。然而,有AI的应用,例如自主驾驶,需要监管。AI还有可能卷入反垄断调查(针对垄断的管制),现在监管者还没想到,但是应该考虑了。

AI as a basic technology should not be regulated. It also seems impractical for the government to stop you from implementing a neural network on your laptop. However, there’re applications of AI, for example autonomous driving, that need regulation. AI also has new implications on anti-trust (regulation of monopolies), that regulators have not yet thought through but should.

关于人工智能法规的讨论,大部分来源于对“有感情的AI”或“邪恶的杀手机器人”的不理性的恐惧,而没有更深入地了解AI到底可以干什么,以及干不了什么。由于今天的AI还不成熟,发展迅速,任何国家的强硬监管都会阻碍该国的AI进展。

Much the discussion on AI regulation stems from irrational fears about “sentient AI” or “evil killer robots,” rather than a deeper understanding of what it can and cannot do. Because AI today is still immature and is developing rapidly, heavy-handed regulation by any country will stunt that country’s AI progress.

然而,有一些人工智能用例需要被监管来保护用户,而且可以加快它们的被接受速度。汽车行业已受到严格监管,以确保安全。仔细思考行业里面监管,如何根据AI的新功能去改变,将有助于整个行业。其他领域也相同,包括制药,军工,金融市场等。但是,监管应该是基于行业特定的,并且基于对用例的深思熟虑,基于我们想达到以及不想出现的后果,而不是去监管基础技术。

However, some AI use cases need regulation both to protect individuals, and to accelerate its adoption. The automotive industry is already heavily regulated to ensure safety. Thinking through how these regulations should change in light of new AI capabilities such as autonomous driving will help the whole industry. Same goes for other areas, including pharmaceuticals, arms control, financial markets, and so on. But the regulation should be industry-specific, and based on a thoughtful understanding of the use-cases and on the outcomes we do/don’t want to see in specific sectors, rather than on the basic technology.

政府也可以发挥重要的作用,帮助解决AI兴起带来的失业,例如提供基本收入和再培训。

The government also has a big role to play to help with the coming job displacements caused by AI, for example to provide basic income and retraining.

最后,AI的兴起正在为一些公司创造了新的途径去竞争,去成为霸主,去击退竞争对手。反垄断监管机构,对竞争的新基础情况的了解远远落在这些企业背后,仍需努力追赶。

Finally, the rise of AI is creating a new ways for companies to compete, become dominant, and shut out competitors. Anti-trust regulators are way behind corporations in understanding this new basis of competition, and have a lot of catching up to do.

共2条回复
hazukijunichi 回复于 2017年09月16日

让我想起群星里

有感情的ai:研究速度+20%。可能触发机器人叛乱的危险科技

堕落爸爸(知识管理者):什么,你在研究ai,不行。你不停止研究就海量舰队打你老家

事实上只要知识管理者天高皇帝远,ai科技一定会研究,毕竟20%的研究速度加成在科技才是爸爸的游戏里非常非常有用。对于现在这个科技才是力量的社会也是同理,zf一旦强行搞事限制其发展,很容易就落后于人

tinyfool 回复于 2017年09月16日

1楼 @hazukijunichi en,没人知道基础科学会带来什么样的连锁反应,所以,任何限制基础研究的行为,都可能搞死自己的未来。但是应用层面确实可以控制一下。

其实基因科学也是一个道理。

登录 或者 注册