Daniel Sinclair @_DanielSinclair Building for young people. Not reading @danielsunread. Lurking behind likes and thinking about social media, communication, & China. Aug. 16, 2019 2 min read

This is... every AI company right now.  https://www.wsj.com/articles/ai-startup-boom-raises-questions-of-exaggerated-tech-savvy-11565775004 

At every level of the tech ecosystem.  https://www.wsj.com/articles/amazon-says-it-has-over-10-000-employees-working-on-alexa-echo-1542138284 

"Amazon staff in areas including Boston, Costa Rica, India, and Romania are listening in to as many as 1,000 audio clips per day during nine-hour shifts."  https://www.zdnet.com/article/amazon-employees-are-listening-in-to-your-conversations-with-alexa/ 

Apple Map's 'black site'  https://www.bloomberg.com/news/features/2019-02-11/apple-black-site-gives-contractors-few-perks-little-security 

"[Apple] added that a very small random subset, less than 1% of daily Siri activations, are used for grading, and those used are typically only a few seconds long." 1% is a whole lot of data and manual labor.  https://www.theguardian.com/technology/2019/jul/26/apple-contractors-regularly-hear-confidential-details-on-siri-recordings 

"The crackdown prompted Bytedance to promise to increase its number of human censors from 6,000 to 10,000 to police and remove 'vulgar' items."  https://www.ft.com/content/bb5b8d5c-a105-11e8-85da-eeb7a9ce36e4 

"[Google is] pledging to change its approach and bring the total number of [moderators to] 10,000 within a year. Included in the 10,000 are many third-party contractors, who also moderate Google’s app store and other Google products."  https://www.washingtonpost.com/technology/2019/08/09/youtubes-arbitrary-standards-stars-keep-making-money-even-after-breaking-rules/?noredirect=on 

Human capital meeting the AI gap is most prominent in China. There is less automation across the entire ecosystem. China companies grow differently, and are human-heavy. This isn't necessarily a bad thing — they're obviously meeting their censorship goals.  https://www.nytimes.com/2018/11/25/business/china-artificial-intelligence-labeling.html 

We would prefer, well, moderation. The China model, supported by cheap labour and infrastructure investment, has been a huge enablement. The U.S. still can't compete at China 'AI' censorship scale — for any of the data we need to tag, transcribe, or interpret. A unique advantage.

So much of the AI is just not real. The tagging and training is still hyper-focused, propped up by a lot of human labour. But, that is actually what we want for some time, at a much larger scale. This is the advantage of the human-enabled ecosystems.

Keyword AI is wrought w/ tremendous fraud and deceit. But we now see what happens when we let 'dumb AI' roam free. Companies should be much more transparent, and we should accept that we are in a human-in-the-loop era of technology — and that's very much what we need & will want.

Give us the human-in-the-loop Siri, not the dumb Siri hiding her weaknesses behind a keyword & singular voice. There is an equitable automation hybrid waiting to be tapped; the most inspiring examples of this aren't coming out of the west. There are better, more sustainable paths


You can follow @_DanielSinclair.



Bookmark

____
Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.

Enjoy Threader? Sign up.

Threader is an independent project created by only two developers. The site gets 500,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Running this space is expensive and time consuming. If you find Threader useful, please consider supporting us to make it a sustainable project.