January 12, 2016

AI is not needed to address Population, Climate Change, Human Development and Education

Google’s chairman Eric Schmidt thinks artificial intelligence will let scientists solve some of the world’s "hard problems," like population growth, climate change, human development, and education.

Rapid development in the field of AI means the technology can help scientists understand the links between cause and effect by sifting through vast quantities of information, said Eric Schmidt, executive chairman of Alphabet Inc., the holding company that owns Google.
“AI will play this role to navigate through this and help us.”

It can also aid companies in designing new, personalized systems. In the future, Schmidt would like to see “Eric and Not-Eric,” he said at a conference in New York, where “Eric” is the flesh-and-blood Schmidt and“not-Eric is this digital thing that helps me.”

Nextbigfuture's position on population growth, climate change, human development, and education is AI is not needed

Larger families in Africa is the remaining aspect of higher population growth. Population going to 11-12 billion by 2100 is because Africa will go from 1 billion to 5 billion. Asia population has flattened other than India and some South Asian countries. The developed world and China have flat to declining population.

A lot of the larger family in Africa issues will be greatly reduced when they have healthcare and public health infrastructure like clean water and sanitation and vaccination. If people have over 99% confidence that their children will die then they will not have 5-6 children to assure that they end up with 2 that grow into adults.

Climate change and CO2 emissions are a unintended byproduct of industrialization. We use coal and fossil fuels for energy and industry and the carbon gets combined with oxygen when it is burned to make CO2. We use about 8 billion tons of coal and about another 4 billion tons of oil. Carbon plus two oxygens means the weight of CO2 is about 40 billion tons. Burning literal mountains of carbons each year also means billions of tons of soot and particulates. Incomplete burning results in this air pollution which also blackens white icecaps and other surfaces. Darkening increases the heat that is absorbed and not reflected. The soot problem results in almost as much heating effect as the CO2 and it also kills more people because of the damage to hearts and lungs. The soot and particulate problem kills over 4 million people each year and makes billions of people less healthy which increases medical costs. The soot and particulate problem is ten to twenty times cheaper to address and can provide faster results of a few years rather than decades.

Again no AI is required to diagnose and prescribe solutions. Clearly when people want to spend many trillions addressing CO2 to get some effect in several decades to a supposedly urgent issue when a lot less money could be spent targeting soot and particulates shows that people are not really trying to solve the big problems.

A 2013 study indicated that the role in climate change for soot is twice as large as previous estimates. Soot has 66% of the impact of carbon dioxide. Mitigating soot would cost about $6 per ton of CO2 equivalent. CO2 mitigation costs about $100 per ton. Nextbigfuture has frequently written that soot is the most cost effective emission target for managing climate. It is also the one with the fastest results. Carbon dioxide mitigation does not impact temperatures for 50-80 years. Fully mitigating soot can also save 1-2 million lives by avoiding the disease from soot pollution

Is AI supposed to make up for human corruption and unwillingness to take obviously better approaches and basic analysis ?

Human development has some big and obvious global problems which are not being addressed.

India and parts of Africa have 40-50% malnutrition. This causes stunting which means smaller people who have some level of brain damage.

There are warehouses of rotting food in India. This is due to corruption where someone made money keeping the food off the market.

For education there are many issues but they are obvious issues and here are some:

* bad cultural concepts that have taken hold in some developed countries where students do not perform an appropriate volume of problem solving.
* school hours have been cut for budget and other reasons which shifts homework responsibility to a significant number of parents who are not prepared, willing or able to help children
* there are uneducated teachers in Africa and undeveloped countries who are illiterate and are unable to properly teach

There are solutions which have work and do work and which do not require AI.

212 Bridge Academies have opened in Kenya during the past four years. Bridge’s “schools in a box” spring up seemingly overnight: In January of 2013, the company launched 51 schools at once, while in September it opened another 78. Bridge now educates roughly 50,000 students in Kenya every day, and its global aspirations may transform the entire project of education for poor youth around the world.

Bridge’s CEO, a former Silicon Valley entrepreneur named Jay Kimmelman, compares his company to Starbucks and McDonald’s — organizations that offer a consistent experience no matter where in the world you encounter them. Beyond its 212 branded academies in Kenya, Bridge has set its sights on Nigeria, Uganda, and India. The founders intend to be serving half a million children in 30 countries by 2015, and 10 million by 2025. “We’ve systematized every aspect of how you run a school,” Kimmelman says. “How you manage it. How you interact with parents. How you teach. How you check on school managers, and how you support them.” And this operational approach gets results. Bridge tests kids six times a year, and a third party performs Early Grade Reading and Math Assessments annually. According to those evaluations, Bridge students are beating out their peers at government and other private schools. In reading fluency, the gap is as high as 205 percent.

The most notable aspect of the Machakos school is that students pay to attend. Since the United Nations made universal enrollment one of its Millennium Development Goals in 2000, governments in Kenya and in much of Africa have eliminated school fees. But outcomes in Africa’s public schools remain deeply disappointing. Dated curricula focus on impractical memorization, and the public school teachers are underpaid and undertrained — spending more energy on managing packed classrooms than on instruction.

Bridge aims to narrow this divide with a radically new take on private school. Tuition is just $5 per pupil per month. There are no student iPads, no science labs. Preschoolers work with clay or blocks; older children learn math with bottle caps and recycled egg crates. Often students write on Dickensian slate boards instead of paper. One of the teachers I met, “Ms. Elizabeth,” completed only high school, and the greatest technological firepower at her disposal was a decidedly analog yardstick.

Bridge invested $15 million in systems development, teacher training, and in-class materials, and built an operations template for every school. (Bill Gates is an investor in the company, as are Khosla Ventures and some other traditional venture capital firms.) All student testing and teacher evaluations would filter through Android-based software monitored by engineers and assessors at Bridge headquarters. Since I met Ms. Elizabeth in 2011, Bridge has equipped teachers and school managers with tablets to manage lessons and track performance. Students take analog tests and teachers upload assessments to a central database. “Accountability is the key,” says Kimmelman. “It’s really weird, and really amazing, and it works.”

For parents hovering around $2 in income per day, a potentially transformative education for their kids was just one of many things they couldn’t afford. The demand, however, remains enormous — the global market for low-cost private education is $51 billion annually. To meet the demand, May says, “we drive the price point low enough so parents can become consumers.”

Bridge’s corporate system allows them to nearly eliminate overhead. The work typically done by secretaries, bursars, principals, and other administrators can be handled remotely, leaving only three nonteaching staff at each school. Parents can transfer fees directly from their mobile phones, an innovation that builds on Kenya’s particularly robust mobile money platforms. (Bridge parents on average make $1.23 per day, but 94 percent of them own mobile phones.) If a teacher fails to open or sync a digital lesson within 15 minutes of its scheduled start, someone from headquarters will take note and call the school to see what’s up. If a teacher is absent, Bridge keeps a paid pool of substitutes standing by. “Our commitment to our parents is that their children will be taught,” May says. “So we invest in, essentially, lots of plan Bs.”

The heart of her pitch is simple: Bridge schools teach kids. Whereas the average government primary school has a 47-to-one student-teacher ratio, Bridge’s is 30-to-one — and it teaches students for two additional hours each day.

The Bridge obsession with consistency and performance produces its most alien attribute: scripted lessons. Because effective lesson plans are a notoriously difficult aspect of teaching, Bridge eliminates any guesswork — dictating classroom instruction down to the noun and to the minute. In Ms. Elizabeth’s subtraction class, she consults the Bridge manual as kids chant and repeat her phrasing with Pavlovian discipline. Her classroom protocol has been written in advance by Bridge’s dedicated curriculum team. This may sound overly doctrinaire, but there are distinct advantages. For teachers, “the examples don’t come off the top of their head, or when they woke up at five in the morning to try and prepare their class,” May says. The scripted approach also allows for incredibly efficient teacher training: Bridge’s seven-week course is lightning-fast compared with traditional accreditation programs.

Форма для связи


Email *

Message *