🔮 Predicting the future with AI / Humans + Tech - #24
+ Zoom update + Tech in the age of coronavirus
Predicting the future has been an obsession of humans ever since we developed the ability for imagination. Shamans, psychics, oracles, clairvoyants, and astrologers have used methods that vary between astrology, palmistry, tarot card reading, summoning spirits, and reading tea leaves, to tell us our future. Every culture through every era has had some form of predicting the future.
In present times, scientists and corporations are attempting to use AI, machine learning, and DNA testing to predict everything from our behaviour, our future lives, our health, and who is likely to commit a crime, among other things.
🤖 AI can’t predict how a child’s life will turn out even with a ton of data
A 15-year study that began in 2000, led by Sara McLanahan of Princeton University, was aimed at understanding how the lives of children born to unmarried parents might turn out over time [MIT Technology Review]. Data from each of the participants in the study were collected when the children were 1, 3, 5, 9, and 15 years old. Sara McLanahan’s team provided this data to hundreds of researchers, computer scientists, statisticians, and computational sociologists. They were asked to submit their best techniques for prediction on six outcomes, that included grade point average at school, their level of “grit,” and the overall level of poverty in their household.
None of them was able to do so with meaningful accuracy. Thankfully!
There are good reasons for predicting these outcomes. Policymakers use the work of social scientists to predict how social outcomes like unemployment and crime will be impacted by policies. An ability to understand how various factors might affect someone’s life will help them implement policies designed to produce the best outcomes.
However, we don’t live in an ideal world. The potential for abuse, even with a 70%-80% accuracy are monumental. What happens when children who are predicted to perform badly in school are denied the same opportunities? What happens if it’s predicted that certain demographics are less likely to perform as well at certain tasks—will that inflame racism, xenophobia, or genderphobia?
🧬 DNA testing for predicting disease risk
In a similar scenario, DNA testing for predicting disease risk is coming soon [The Conversation]. The positive side is that those who find out they are at risk for certain diseases can take precautions to prevent or delay these diseases. On the other hand, there are serious ethical, privacy, and regulatory concerns. Who has the right to your DNA test results? Will insurance companies deny you cover or raise your premiums based on your risk levels? Will they mandate that you take a test before they insure you? What about those who would rather not know and live life like we’ve been doing for centuries?
Genetic discrimination is already a reality in Australian Life Insurance [National Center for Biotechnology Information]. With the plummeting cost of DNA testing, it’s not long before it spreads to all regions worldwide. We need to think deeply about these issues and how they will impact society before deploying them.
👮♂️ Minority Report is now a reality
🇬🇧Police in the UK are piloting a system using AI to identify future criminals [The Next Web]. The system predicts which low-level offenders are likely to commit more high-level crimes in the future. This information is then used to inform interventions such as jobs training, mental health support, and substance abuse treatments that would reduce the risk of them committing crimes in the future.
UK government’s Centre for Data Ethics commissioned a report on the use of AI by the British Police. The report revealed major legal and ethical concerns over the growing use of data analytics by police and said that official national guidance on the use of algorithms in policing was urgently required.
🇰🇷 Meanwhile, South Korea is also deploying Crime-Predicting AI cameras in Seoul [Global Government Forum]. The system will be able to detect whether people are carrying dangerous objects, suspicious items such as large bags, and could also determine whether someone is being followed.
🇨🇳 China has also deployed AI crime-prediction and social control in Xinjiang [Express]. The system identifies people who exhibit abnormal behaviour based on where they go and what they do.
Preventing crime is great. But AI and machine learning are based on assessing risk using statistical analysis. Using risk models that are not 100% accurate to target potential criminals, assuming they will be guilty in the future, and making them undergo intervention programs is unethical in my opinion and may even border on human rights violations.
😯 You want a surprise
Predictions are fun to make when there are large amounts of uncertainty as to whether they are true. Because then, we don’t make decisions based on those predictions, other than placing a few bets. One of my favourite philosophers, Alan Watts, says:
“A completely predictable future is already the past. You’ve had it! That’s not what you want. You want a surprise.”
Think about sports. If we knew the result of a game or competition beforehand, most of the joy of watching sports is gone. We always want a surprise.
Let’s leave other parts of our life to surprise us too. Surprises are what make life interesting.
You can hear Alan Watt’s quote in his short talk on “What do you desire?” [YouTube 6:39].
When in doubt, listen to Alan Watts. He has immense amounts of wisdom. Some of his concepts are hard to grasp initially but the more you think about them, the more they make absolute sense.
🖥 Zoom update
Since last week’s newsletter on Zoom’s privacy and security shortcomings, there have been some new developments. Several organizations including Google, SpaceX, NASA, New York City’s Department of Education, Taiwan’s Government, Germany’s Foreign Ministry, and the Australian Defense Force among others, have all banned their employees from using Zoom [TechRepublic].
To be safer, you can use Zoom via the browser without installing the app. This is a little more secure, except that they’ve hidden that functionality. Learn how to do so here [TechCrunch].
Zoom has addressed some of its privacy and security issues in an update released a few days ago [TechRadar].
🦠 Tech in the age of coronavirus
📇 Contact tracing and tracking apps are all the rage these days. In an unprecedented collaboration, Apple and Google are partnering to launch a joint COVID-19 contact tracing tool for iOS and Android [TechCrunch]. First, they will release an API that health agencies can integrate into their apps. The second and more important phase will be building the contact tracing into their respective operating systems. They say the process will be opt-in.
Bluetooth technology will enable them to trace every other phone that is within range. In the event that one person is declared COVID positive and shares that information on this system, all other people that the person came into contact with over the last 14 days will be automatically alerted via their phones with information on what to do next. Both Apple and Google say that this system will be built with privacy in mind, and no personally identifiable information will be used or stored.
With Apple and Google having over 95% of the market share for phones via iOS and Android, they are in the best position to make this type of large-scale contact tracking and tracing possible. It’s great to see both companies collaborate and work together to help provide solutions with privacy in mind.
📍 Pinterest’s CEO, Ben Silbermann, is also building a COVID-19 tracking app in collaboration with a team of top scientists [TechCrunch]. It is called “How We Feel” and makes it easy for people to self-report whether or not they feel well and what symptoms they’re experiencing. The goal is to make it quick and easy to report. They also insist that they have designed the whole experience with privacy in mind and that people’s information is used responsibly.
Silbermann and his wife Divya will also be providing a donated meal to nonprofit Feeding America for every time a person downloads and uses the app for the first time, up to a maximum of 10 million meals.
Quote of the week
"Prophesy is a good line of business, but it is full of risks. "
—Mark Twain in Following the Equator
I wish you a brilliant day ahead :)