我们不是邪恶的

AI is driven by a joint worldwide effort of programmers, data scientists, and mathematicians. Overall, this particular global subculture, to which I count myself, is mostly optimistic, believing in progress, and ultimately wanting to improve the world through technology.

人工智能是由程序员，数据科学家和数学家在全球范围内共同努力的结果。 总体而言，我认为自己是这种特殊的全球亚文化，大多数人是乐观的，相信进步，并最终希望通过技术改善世界。

Following the news, you might get a different picture of AI and the people behind it. When the Banjo story broke [1], we learned that major figures throughout AI history have been White supremacists [2]. Since the COMPAS scandal [3] we read about gender and racial bias in just about any major AI system out there. Just a few days ago, IBM, Amazon, and Google desperately pulled the plug on their facial recognition Softwares [4–6], seeing no other way to prevent unfair treatment of minorities. Even worse, statistics, the very mathematical discipline behind all modern Machine Learning and Data Science, has now itself been suspected of being racist, as it has been developed, in large parts, by eugenicists [7, 8].

新闻发布后，您可能会对AI及其背后的人有不同的了解。 当班卓琴的故事破裂[1]时，我们了解到整个AI历史上的主要人物都是白人至上主义者[2]。 自从COMPAS丑闻[3]以来，我们几乎了解了几乎所有主要的AI系统中的性别和种族偏见。 就在几天前，IBM，亚马逊和谷歌拼命拔掉了他们的面部识别软件[4-6]，发现没有其他方法可以防止对少数民族的不公平待遇。 更糟糕的是，统计学是所有现代机器学习和数据科学背后的非常数学的学科，由于优生主义者在很大程度上发展了统计学，所以它本身现在已经被怀疑是种族主义者[7，8]。

Although statistics, like science in general, is certainly full of ethical contamination (to use a term by Ron Howard [9]), it is, by itself, hardly evil. Neither are we, the vast majority of passionate data scientists and Machine Learning engineers from all around the world, who rely on statistics in everything we do.

尽管统计数据像一般科学一样，当然充满了伦理污染(使用罗恩·霍华德[9]的术语)，但就其本身而言，它几乎不是邪恶的。 我们都不是，来自世界各地的绝大多数热情数据科学家和机器学习工程师，他们在所做的一切事情中都依赖统计数据。

No, statistics is not the problem, it is vital to the solution. It is the one necessary tool to reveal discrimination and injustice anywhere.Only by means of systematic data analysis and statistics, can the unfair practices prevalent in today’s companies and institutions be at all revealed. Statistics allows us to quantify injustice and to monitor whether we are making progress reducing it.

不，统计不是问题，这对解决方案至关重要。 它是在任何地方揭示歧视和不公正现象的必要工具。只有通过系统的数据分析和统计，才能揭露当今公司和机构中普遍存在的不公平做法。 统计数据使我们能够量化不公正现象，并监测我们是否在减少不公正方面取得了进展。

How did ProPublica uncover the COMPAS scandal? By use of statistics and Data Science, as you can see from their code. How do we know that AI research and development has a diversity problem? By looking at — you guessed it — statistics [10]. How could Cathy O’Neil write “Weapons of Math Destruction”? By drawing on her knowledge and experience as a mathematician and Data Scientist. She’s one of the many of us who actually care.

ProPublica如何揭露COMPAS丑闻？ 通过使用统计和数据科学， 您可以从他们的代码中看到 。 我们如何知道AI研究与开发存在多样性问题？ 通过查看(您猜对了)统计数据[10]。 凯茜·奥尼尔(Cathy O'Neil)怎么写“数学毁灭武器”？ 通过利用她作为数学家和数据科学家的知识和经验。 她是我们当中真正关心的许多人之一。

The problem is not the AI that learned your unethical behavior. The problem is that you behaved unethically, in the first place!

问题不在于学会了您的不道德行为的AI。 问题是，首先，您的行为举止不道德！

True, incidents such as the COMPAS scandal show that AI makes unethical decisions. But, dear companies and institutions, the problem is not the AI that learned your unethical behavior. The problem is that you behaved unethically, in the first place!

没错，COMPAS丑闻等事件表明AI做出了不道德的决定。 但是，亲爱的公司和机构，问题不在于学会了您的不道德行为的AI。 问题是，首先，您的行为举止不道德！

A lot of work is done, to identify and correct this. Ethical AI has become an active field of research, resulting in new tools, methods, and best practices to achieve data transparency and algorithmic explainability. And yes, the era of explainable AI is going to be uncomfortable for some corporations and institutions because with increased transparency comes additional scrutiny. Ultimately, though, there is no alternative. It will be through the contribution of statistics that inequality is detected, measured, and fought systematically.

为了确定并纠正此问题，需要做很多工作。 伦理AI已成为一个活跃的研究领域，它产生了新的工具，方法和最佳实践，以实现数据透明性和算法可解释性。 是的，对于某些公司和机构来说，可解释的AI时代将是令人不舒服的，因为随着透明度的提高，将进行更多的审查。 但最终，别无选择。 依靠统计的贡献，系统可以发现，衡量和消除不平等。

And we, you and I, will be the ones to bring this about.

而我们(您和我)将成为实现这一目标的人。