independent news and opnion

Artificial Intelligence Is Failing Humans

0 24

In a recent New York Times editorial titled “Why We Should Stop Fetishizing Privacy,” entrepreneur Heidi Messer posited the ultimate list of techno-optimist clichés. Tech companies drive the economy, bringing health, wealth, jobs, and truth. Those who caution against a host of risks such as monopoly, hubris, and shortsightedness should be dismissed as ignorant “privacy evangelists.” Public regulation is bad because tech companies have “the talent and resources” to protect us against cyberwarfare and “foreign and criminal intrusion.” We should follow the example of “digital natives,” who “start with an awareness that their data isn’t private.” According to Messer, public oversight would only gum up the workings of all the utopian delights of this shiny new world. Given such blessings, why would we want to break up big tech companies?

There is much to unpack in Messer’s breezy dismissal of both unchecked monopoly power and the invasive apparatus of totalistic technological surveillance. For present purposes, however, I would like to think about how the concept of the self is affected by the widening use of algorithms to translate more and more bits of ourselves into numerical representations. Artist Trevor Paglen gives a succinct example: When people upload pictures of their kids, algorithms reading those photos feed invisible data sets in ways that may eventually influence something as apparently unrelated as those children’s health insurance. Similarly, if a teenager uploads a picture of herself having a beer, her underage drinking may be marked as information that can be sold, utilized by police departments whose scrutiny “will be guided by your ‘pattern of life’ signature,” warns Paglen. “When you put an image on Facebook or other social media, you’re feeding an array of immensely powerful artificial systems information about how to identify people and how to recognize places and objects, habits and preferences, race, class, and gender identifications, economic statuses, and much more.”

Recently, a 10-year-old in Maryland shared clearly marked play money with classmates while riding on his school bus. The driver contacted his supervisor. Police were called, and finally, the Secret Service—all to investigate the child for counterfeiting. While this is absurd on its face (and yes, the child was black), what’s more invisibly sad is that each time a person enters a database for having had contact with police, it will affect all sorts of other life chances, including risk assessments for employment, credit, and child custody.

That’s largely because artificial intelligence dispenses predictive computations based only on what it is trained—by humans—to see. Many universities now use Canvas, a course management platform on which students can discuss material or share lecture notes while their instructor is talking. When I was being trained to use the program, I noticed that the IT department had a screen on which the entire faculty was numerically ranked based on who generated the most comments during lectures. I was told that it would help us know what parts of a lecture stirred interest, but to me, it seemed only a whisker away from a Kim Kardashian standard of generated buzz as professorial achievement—a test I surely fail because I often tell my students to close their laptops.

Source link

You might also like

Thanks !

Thanks for sharing this, you are awesome !