Artwork

内容由Video Archive – The Conference by Media Evolution提供。所有播客内容(包括剧集、图形和播客描述)均由 Video Archive – The Conference by Media Evolution 或其播客平台合作伙伴直接上传和提供。如果您认为有人在未经您许可的情况下使用您的受版权保护的作品,您可以按照此处概述的流程进行操作https://zh.player.fm/legal
Player FM -播客应用
使用Player FM应用程序离线!

Ovetta Sampson - Design Principles for a Pluralist Automated Future

19:16
 
分享
 

Manage episode 375847923 series 1191096
内容由Video Archive – The Conference by Media Evolution提供。所有播客内容(包括剧集、图形和播客描述)均由 Video Archive – The Conference by Media Evolution 或其播客平台合作伙伴直接上传和提供。如果您认为有人在未经您许可的情况下使用您的受版权保护的作品,您可以按照此处概述的流程进行操作https://zh.player.fm/legal

"Data is the love language of machine learning, but we must remember that it is not true."


We all create data. And all data is created by people. Ovetta Sampson wants us to remember this, both in order to centre humanity but also to clarify the vulnerabilities of data. We are biased, so the data we create is infused with biases as well. Whether it is by the sin of omission or the use of inequitable variables, traumatised datasets manifest in real world situations such as applying for a bank loan or decisions made on housing and education.


Ovetta urges particular caution for the encounters between humans and machines in the era of AI and machine learning. It's not Skynet, not yet, but ceding decision making responsibility to such systems can lead to harmful consequences. The best way of countering these resides in responsible, human-centred design frameworks which capture the minimum viable data, maintain balance in the exchange of what people give and what they receive, and include iterative privacy by default.


Ovetta ends with a rigorous set of responsible design practices to combat the amplification of our human biases by AI systems.


  continue reading

500集单集

Artwork
icon分享
 
Manage episode 375847923 series 1191096
内容由Video Archive – The Conference by Media Evolution提供。所有播客内容(包括剧集、图形和播客描述)均由 Video Archive – The Conference by Media Evolution 或其播客平台合作伙伴直接上传和提供。如果您认为有人在未经您许可的情况下使用您的受版权保护的作品,您可以按照此处概述的流程进行操作https://zh.player.fm/legal

"Data is the love language of machine learning, but we must remember that it is not true."


We all create data. And all data is created by people. Ovetta Sampson wants us to remember this, both in order to centre humanity but also to clarify the vulnerabilities of data. We are biased, so the data we create is infused with biases as well. Whether it is by the sin of omission or the use of inequitable variables, traumatised datasets manifest in real world situations such as applying for a bank loan or decisions made on housing and education.


Ovetta urges particular caution for the encounters between humans and machines in the era of AI and machine learning. It's not Skynet, not yet, but ceding decision making responsibility to such systems can lead to harmful consequences. The best way of countering these resides in responsible, human-centred design frameworks which capture the minimum viable data, maintain balance in the exchange of what people give and what they receive, and include iterative privacy by default.


Ovetta ends with a rigorous set of responsible design practices to combat the amplification of our human biases by AI systems.


  continue reading

500集单集

所有剧集

×
 
Loading …

欢迎使用Player FM

Player FM正在网上搜索高质量的播客,以便您现在享受。它是最好的播客应用程序,适用于安卓、iPhone和网络。注册以跨设备同步订阅。

 

快速参考指南