weekly column
Each week, find a commentary on something connected to verses of Torah or another source of wisdom
|
Each week, find a commentary on something connected to verses of Torah or another source of wisdom
|
The Last of Deuteronomy The secret things belong to the Lord our God, but the things that are revealed belong to us and to our children forever, that we may do all the words of this law. Deuteronomy 29:28 Algorithms are sequential mathematical calculations that can arrange and rearrange all sorts of data mostly used in problem-solving by computer. If you understand that, you are ahead of me. You are familiar with algorithms because they determine what turns up in your social media feed – advertisements, click-bait, postings by friends (that is, not necessarily personal friends, but people with whom you have a shared algorithmic connection). If you watched a video of cute little kittens, you will see other cat videos recommended to you. If you ordered a package of warm socks for the winter, you will hear from many footwear companies and apparel purveyors. And if you are reading this column on an electronic device, do not be surprised, when you next log into YouTube, near the top of your list “chosen for you” is a clip of a black-and-white game show from the late 1950s starring Garry Moore. Algorithms are not magic. They are written by programmers, one of whom was probably the kid who programmed your VCR or hooked up your desktop back in the day. The concept is very simple. By sorting and resorting little bits of information, these calculations can provide an approximation of the answer to any quantifiable question. “What will Jack watch on Netflix” is more accurately predicted by figuring out what Jack has already watched on Netflix, which the good folks at Netflix know because there is an algorithm that tracks what Jack watches. They are the basis of artificial intelligence (AI). And that means that they have application in all sorts of fields. For example, if you have ever used a web-based service to find candidates for employment, they have been filtered by algorithms. Especially in small companies, the ability to sort applications by appropriate criteria can be invaluable. Geography, age, education, experience – all of these variables can be considered by algorithms to reduce hundreds of applications to the appropriate few. Likewise, a consumer seeking a line of credit or a mortgage can qualify or not based on objective criteria. Or, college admissions can be evaluated on a schools standards for grades, test scores and desire for a diversified student body. There are those who make the argument that such mathematically precise calculations help to eliminate the biases that reside in the hearts and minds of decision-makers. But unconscious bias resides in the heads and hearts of all human beings, including the creators of algorithms. With perhaps the very best of intentions, the inclusion of a zip code, a marker of illness, previous employment in a troubled company may disadvantage the applicant for financing a new car, getting favorable rates on health insurance or being considered for employment. The company may be unaware; the applicant may be unaware; the programmer may be unaware. The human aspect is missing. Well-targeted questions, opportunities for nuanced discussions, deepening the understanding of past circumstances – all of them hard to quantify and therefore considered outside the realm of the computations – cannot mitigate a circumstance that does not get past the first threshold. A recent anecdote in the real estate section of the newspaper illustrated the phenomenon. An interracial couple wanted to sell their house near the border between their mostly-White neighborhood and a mostly-Black neighborhood. The husband, Black, was home alone when the appraiser came to value their property, and by comparing it to the houses owned by neighboring Black homeowners, offered a valuation much lower than the couple expected. The wife, White, arranged for a second appraisal and greeted the appraiser. The valuation went up $150,000 – about 35%. And my interest in the occasional order of kosher pastrami does not prevent the company that ships it to me from offering me decidedly non-kosher selections. The solution is not simple. Ethicists need to look at algorithms. Programmers must work among diverse colleagues. Watchmen need to watch the watchmen. The secret things in our hearts and minds may be unknown to us and to our children. Perhaps it is not only the Holy One who can ferret them out.
0 Comments
Leave a Reply. |
Archives
October 2023
Categories |