Thread by @ram_ssk

To celebrate @JasmineMcNealy leading the Marion B. Brechner First Amendment Project, I watched her awesome talk on "Ecological Approach to Data" @datasociety

TL;DR: Property analogies fail us when we talk about data

Notes⬇️(esp for @AndrewYang) 1/x

It is MY data. It is something I own, right? And something, I own is property, right?

Turns out property is a creation that is protected by the state. This allows ownership of a "thing" This could be tangible like my house or intangible like copyright 2/

Here is why Property as an analogy suck
1) Historically, "property" has never been inclusive. For instance,
- African Americans were treated like property and denied self fulfilment.
- Women were denied ownership of property and self. 3/

2) Eminent Domain

The US Constitution gives Eminent domain resides -- so, you can have your property taken from you by the State (with 'just compensation')!

(side note: @elizabeth_joh wonderfully narrates the Takings Clause - trumpconlaw.com/8-the-takings-…) 4/

3) Regulatory Take In -- You may own the property but it is sooooo heavily regulated that it is not worth it

5/

4) Rule of Capture -- For data to be useful, it needs to be "captured" through donation or unwittingly. In property law, you have "Rule of Capture", which she explains as first to capture possess it. You may have designed to capture, but until you actually capture, not yours. 6/

Here is how Jasmine explains it: If your property and your neighbor's property are sitting on an oil field, then if you drill first and capture all the oil, you can suck it all up without asking your neighbor’s permission.

The first company to suck up your data owns it 7/

5) Property does not take into account the "Networked nature of data"

The golden state killer did not participate in GEDMatch. He was caught through his cousins. They did not find him through HIS data.

sciencemag.org/news/2018/10/w…

8/

This is where Jasmine says, property rights do not take network/attachments into account.

**Data always comes with attachments – bias, labels, how to process it, context**

(Sidenote: Can I tattoo this?) 9/

So, in order to account for the networked nature of data, she proposes "an ecological approach" wherein we account that the data is placed in increasingly broader context and then rationaled. 10/

Some fun side notes:

1) Jasmine pointed how SPSS from @IBMAnalytics has default settings -- 0 for men; 0 for white. Everything else is perceived as "deviations". 11/

2) In her panel, Sareeta Amrute expertly remarks out how Jasmine is flipping the conversation - when there is a huge push to simplify things, Jasmine's ecological approach is asking to point out the complexities. 12/

Congratulations @JasmineMcNealy!! We are lucky to have your scholarship!