Kartik Hosanagar’s A Human’s Guide to Machine Intelligence is a phenomenal book that notes how algorithms and artificial intelligence are shaping our lives, and what can one do to stay in control. As they are embedded in every popular tech platform and every web enabled device, these algorithms and artificial intelligence carry out a plethora of functions for us, from choosing what products we buy to how we find a job.
Kartik Hosanangar through his book tries to explain how and why we need to arm ourselves with a better, deeper and a more nuanced understanding of the phenomenon of algorithmic thinking. He examines various episodes of such algorithms going rogue and why one needs to be more cautious while using such technology.
Here are some facts about AI from the book!
Match.com, one of the most popular dating website in the United States was launched in 1995 and aimed at finding the perfect partner for people. However, in 2011, a Financial Times reporter exposed that although the company’s algorithm asked people to list the characteristics they would want in an ideal partner, these lists were ignored. Rather, the people that the website urged the users to reach out to, was based on the profiles the users had visited previously.
“The conventional narrative is that algorithms will make faster and better decisions for all of us, leaving us with more time for family and leisure. But the reality isn’t so simple.”
∼
The feature of autocomplete on Google, which was first introduced by Kevin Gibbs, is something that we now take for granted. There have been many instances where this feature has proved to reiterate the prejudices that are assumed regarding certain subjects.
“But it’s far more disturbing to ask if Google might have unintentionally led impressionable people who did not initially seek this information to webpages filled with biased and prejudiced commentaries, effectively delivering new audiences directly to hate-mongering sites.”
∼
The algorithms used by Netflix, Amazon, and other online firms through collaborative filtering produce a biased range of shows or products that are popular, rather than promoting obscure and niche items. This is primarily because the algorithms of these online firms tend to recommend things based on what others are consuming.
“We developed simulations of several commonly used recommendation algorithms to test the theory, and they indeed demonstrated that these algorithms can create a rich-get-richer effect for popular items.”
∼
Following the introduction of Google’s famously talked about ranking algorithm, which was made public in the year 1999, it resulted in various website owners creating “shadow” websites which would link back the users to their primary domain. Similarly, in the present age, Instagram and Twitter are working hard to minimise the presence of bot and spam accounts that are made to like and repost other accounts, thereby boosting the spammers’ rank on the platforms’ ranking algorithm.
∼
“And manipulability will only become an increasing concern as algorithms come to be used in other domains with more serious consequences. Suppose a fraudster knew exactly what rules credit card companies used to flag suspicious activity, or a terrorist knew exactly what TSA screening systems were looking for in their image-processing algorithms. With that knowledge, it would become easy to avoid detection.”
∼
Various social media websites such as Facebook, Twitter and also search engines such as Google have become a great source of information and news for people over a period of time. However, concerns over the use of personalization algorithms have come to grab the attention of many, as the algorithms of such tech companies access information about our preferences over time, creating a “filter bubble” which only shows things that relate to our preferences. This results in the barring of alternate perspectives.
“As we engineer our algorithmic systems, the algorithms themselves certainly deserve a high degree of scrutiny.”
A Human’s Guide to Machine Intelligence is an entertaining and provocative look at one of the most important developments of our time and is a practical user’s guide to this first wave of practical artificial intelligence.
