How Can A.I. Improve People's Lives?
RCSU Science Challenge 2019 Submission
Artificial intelligence (A.I) is [1] the simulation of human intelligence processes by machines leading to the machine being able to learn, reason and apply their knowledge. 2018 was truly a watershed year for A.I, where we saw a myriad of new technologies and ideas unveiled ranging from [2] brain-like computing chips containing artificial neurons which allow the chips to learn information for the evaluation of patterns in large data sets to extensive facial recognition systems which are actively being used to success in [3] crime detection and forensics by the FBI. Not to mention the growth in the number of start-up companies ([4] in the US there has been an increase in 23% from 2017 to 2018) alone where investors are pouring [5] billions to transform entire industries with A.I. for example UiPath, which is developing robotic automation to optimize business processes.
[6]Worldwide spending on A.I. systems is set to reach $35.8 billion during 2019 which is a staggering increase of 44% since 2018, and business value derived from A.I. is forecast to grow from $1.2 trillion in 2018 to $2.1 trillion by the end of 2019 - it is fair to say that the current state of A.I. is organic (i.e. constantly changing) due to corporations such as Google (e.g. with DeepMind, a world leader in AI research), Amazon, Microsoft and Amazon getting increasingly involved despite the relative youth of A.I, but to what extent is this directly improving people's lives and welfare? Sure, reciprocal advantages are present which arise from the industries developing, but for artificial intelligence to truly have a profound impact on people’s lives in the future, then what needs to be targeted, advanced and revolutionised (alongside A.I) are the primary sectors of society. Predominantly, these include the food and energy industry and the education sector.
I believe that if a majority of research and funding goes into these areas, a positive-multiplier-effect (PME) will unfold which will in turn indirectly benefit other sectors. What I mean by a PME in this context is essentially a domino effect, but each effect is beneficial and causes another beneficial event which in turn causes another and so on. An example of a plausible PME would be, that if we revolutionised the education system making it virtually available and accessible to all, a flood of new creative ideas would emerge, which if implemented, would bring a whole new level of innovation leading to unforeseen positives, furthering improvements in quality of life.
An example of this would be with Blue River Technology, which is a company that is transforming weed control by using drones that ‘precision spray’ - in a world where herbicide resistance is a growing concern [7] with over $25 billion spent each year controlling it and 3 billion pounds of herbicides deployed annually, it is fair to say that fighting these hurts profitability within this sector hugely ([8] annual losses are estimated at $43 billion). The ‘precision spray’ drone has successfully managed to decrease herbicide costs by a staggering 90% to farmers who have acquired this system.
Building on from this idea, if an A.I. could be developed where it can learn and cater for the needs of each and every type of plant in a farmer’s field, we would see a dramatic increase in sustainability and quality of produce. Today, the best practice is to treat all plants as if they have the same needs, but this paradigm needs to be changed to make every individual plant count at scale according to Blue River Technology.
It would be like each plant having a personal carer, but this carer can adapt and change itself to cater for every single plant in the field. Combining an A.I. like this, with already existing technologies such as drones which can monitor soil temp, H2O levels and topography (shape and features of fields) may be the key to unlocking a near-100% yield. Current steps taken to reach this target have mainly included the adoption of genetically modified (GM) crops e.g.[9] by using GM corn, their yield relative to non GM equivalents have increased nearly 5 fold from 5.6 to 24.5% and also the development and distribution of more advanced soil fertility kits to ensure that the farmers’ soils are not lacking in macronutrients (nitrogen, phosphorus and potassium) because a deficiency in these will lead to hampered growth and low yields since nitrogen and phosphorus are the main constituents of proteins and nucleic acids and potassium plays an important role in regulating CO2 uptake for photosynthesis.
This would affect a multitude of lives – by fundamentally increasing the supply, prices for produce should be driven down over time which would make it much more accessible to everyone, even the least privileged in our world. The PME caused from a healthy, nutritious diet (which is currently inaccessible/unavailable for 25-30% of the world’s population [[10] the Grocery Gap]) would be extraordinary, exponentially increasing productivity levels and stimulating the economy of low-income-countries, not to mention how it may break many out of a vicious poverty cycle, by allowing them to work more efficiently. Of course this impact will not be instant: prices for produce may not change initially due to the expense of the technology, and distribution of food to where it’s needed the most has always been a difficult problem to solve for reasons such as [11] the lack of markets, inadequacy of transportation to the markets, and the inability to afford production costs.
In making this idea a reality, botanists/phytologists would need to conduct further research into plant biology. Specifically, they would need to note chemical and molecular signs exhibited by the plant which indicate whether it is in need of a specific mineral ion before visible signs are shown e.g. chlorosis (yellowing of plant leaves due to iron and magnesium deficiencies) in order to completely negate the effects of the symptom. Chlorosis would stunt a plants growth for a reasonable duration of time by inhibiting photosynthesis if it goes to full effect. Visible signs are a response to something chemically - if that chemical signal can be detected and rectified at an early stage by the drone, it would be like there was never a deficiency in the first place (similar to how if a cancer is detected earlier, there is drastically higher chances of successful treatment).
The build-up of atmospheric carbon dioxide contributes towards climate change because the polarity of the compound means the bonds within it can absorb infra-red radiation emitted from the Earth, thus heating it up, and carbon soot causes global dimming by reflecting sunlight back into space and result in cooler than expected temperatures.
Not to mention the sustainment of the fossil fuels for future generations resulting from the employment A.I. in this way. Energy security is about long-term sustainability, not just about servicing today’s demand by exhausting supplies.
In the future, I think that schools and learning establishments will implement A.I. in the light of using it for personalised learning. [15] Every student has a different learning style, visual, aural, verbal etc, but not every style is catered for, leaving some students not being able to maximise their full potential which is not ideal in a world which is becoming dominated by tertiary and quaternary industries (essentially ones that are very knowledge, information and skill driven).
The A.I would provide and assist the teacher with information in multiple formats to improve the students experience and efficacy of learning. This same A.I can be reprogrammed into a ‘chat bot’ perhaps, and acting as a portable, personalized tutor accessible from any device. Multiple A.I chat bots have already been coded for recreational use e.g. [16] Cleverbot, which is constantly learning from its users. I think something similar, but in terms of providing educational, tailored resources and guidance to students of a certain learning style can be achieved.
For example, let’s say you have an aural learner, the A.I could instantaneously combine a series of pitches, rhythms and specific subject content to create a song, making the content more understandable, memorable and applicable.
A 24/7, 365 days a year support network consisting of A.I chat bots/tutors, which would be personalized to individual students? Seems quite far fetched, but the [17] Georgia Institute of Technology are already using an A.I. teaching assistant, named Jill Watson which has been reported to optimize learning and reduce the workload of Ph.D. students, whilst also adopting a colloquial tone to seem almost real. [18] Ashok Goel (the developer of Jill Watson) states that his own students did not realise that their teaching assistant was not human which is important in the sense that even with an A.I machine, human-like interaction can still be simulated.
This shows that the foundations are present, we just need to develop onto what we have and expand the usability to reach and affect as many students as possible. The questions answered usually have firm and objective solutions, but Ashok Goel sees this as the basis of a start-up in providing lucrative tools in the education field and
Not only would an A.I. of this sort cost an incredible sum ([19] building one chat bot can cost anywhere between $40,000 to $100,000 depending on complexity) but logistically would be hard to implement. [18] Ashok Goel had to collate four semesters worth of data, including over 70,000 questions and answers: only then could he start ‘training’ it and tailor it to his specific course – a very lengthy process indeed. Even then, Jill Watson was not an instant success, because initial testing from Ashok showed that she was giving incorrect information and even to this day she is only 97% accurate.
To make this somewhat of a wide scale system, strict trials should take place, showing off the benefits from using this system. Starting off with a few learning establishments – students within it can be divided into two groups, one using the A.I. alongside their teacher and one without - their performance can be monitored to draw conclusions. Even though they are not directly linked to knowledge, cognitive function tests devised by neuroscientists could also be used to show if A.I. based learning has any other advantages.
[20] Louis Columbus, a contributor to Forbes claims that “Artificial Intelligence will enable 38% financial gains by 2035,” but if these ideas are implemented, I believe that it will direct A.I to be an improving force (in terms of quality of life), rather than merely a force for profit and financial gain.
The profit that I am referring to, which is not necessarily monetary includes something like acquiring a skilled, educated workforce who have nutritious diets to increase their productivity. This would be seen as tremendous ‘profit’ to a newly emerging economy or low-income country because it would ideally stimulate their economic development. The monetary profit would arise as the result of the improving force.
[6]Worldwide spending on A.I. systems is set to reach $35.8 billion during 2019 which is a staggering increase of 44% since 2018, and business value derived from A.I. is forecast to grow from $1.2 trillion in 2018 to $2.1 trillion by the end of 2019 - it is fair to say that the current state of A.I. is organic (i.e. constantly changing) due to corporations such as Google (e.g. with DeepMind, a world leader in AI research), Amazon, Microsoft and Amazon getting increasingly involved despite the relative youth of A.I, but to what extent is this directly improving people's lives and welfare? Sure, reciprocal advantages are present which arise from the industries developing, but for artificial intelligence to truly have a profound impact on people’s lives in the future, then what needs to be targeted, advanced and revolutionised (alongside A.I) are the primary sectors of society. Predominantly, these include the food and energy industry and the education sector.
I believe that if a majority of research and funding goes into these areas, a positive-multiplier-effect (PME) will unfold which will in turn indirectly benefit other sectors. What I mean by a PME in this context is essentially a domino effect, but each effect is beneficial and causes another beneficial event which in turn causes another and so on. An example of a plausible PME would be, that if we revolutionised the education system making it virtually available and accessible to all, a flood of new creative ideas would emerge, which if implemented, would bring a whole new level of innovation leading to unforeseen positives, furthering improvements in quality of life.
The
Farming Industry
A.I. has
already been implemented within many areas of farming, such as with the use of
agricultural drones which overtime, learn and adapt to the environment they are
in and suggest realistic modifications to each and every acre to promote the
highest yields and least amount of dissipation (waste).
 |
Figure 1: Typical agricultural drone.
They are not dissimilar to the
drones used in photo/videography
|
Building on from this idea, if an A.I. could be developed where it can learn and cater for the needs of each and every type of plant in a farmer’s field, we would see a dramatic increase in sustainability and quality of produce. Today, the best practice is to treat all plants as if they have the same needs, but this paradigm needs to be changed to make every individual plant count at scale according to Blue River Technology.
It would be like each plant having a personal carer, but this carer can adapt and change itself to cater for every single plant in the field. Combining an A.I. like this, with already existing technologies such as drones which can monitor soil temp, H2O levels and topography (shape and features of fields) may be the key to unlocking a near-100% yield. Current steps taken to reach this target have mainly included the adoption of genetically modified (GM) crops e.g.[9] by using GM corn, their yield relative to non GM equivalents have increased nearly 5 fold from 5.6 to 24.5% and also the development and distribution of more advanced soil fertility kits to ensure that the farmers’ soils are not lacking in macronutrients (nitrogen, phosphorus and potassium) because a deficiency in these will lead to hampered growth and low yields since nitrogen and phosphorus are the main constituents of proteins and nucleic acids and potassium plays an important role in regulating CO2 uptake for photosynthesis.
This would affect a multitude of lives – by fundamentally increasing the supply, prices for produce should be driven down over time which would make it much more accessible to everyone, even the least privileged in our world. The PME caused from a healthy, nutritious diet (which is currently inaccessible/unavailable for 25-30% of the world’s population [[10] the Grocery Gap]) would be extraordinary, exponentially increasing productivity levels and stimulating the economy of low-income-countries, not to mention how it may break many out of a vicious poverty cycle, by allowing them to work more efficiently. Of course this impact will not be instant: prices for produce may not change initially due to the expense of the technology, and distribution of food to where it’s needed the most has always been a difficult problem to solve for reasons such as [11] the lack of markets, inadequacy of transportation to the markets, and the inability to afford production costs.
 |
Figure 2: The cycle of poverty, also referred to as the spiral of
decline.
|
In making this idea a reality, botanists/phytologists would need to conduct further research into plant biology. Specifically, they would need to note chemical and molecular signs exhibited by the plant which indicate whether it is in need of a specific mineral ion before visible signs are shown e.g. chlorosis (yellowing of plant leaves due to iron and magnesium deficiencies) in order to completely negate the effects of the symptom. Chlorosis would stunt a plants growth for a reasonable duration of time by inhibiting photosynthesis if it goes to full effect. Visible signs are a response to something chemically - if that chemical signal can be detected and rectified at an early stage by the drone, it would be like there was never a deficiency in the first place (similar to how if a cancer is detected earlier, there is drastically higher chances of successful treatment).
The Energy Industry
A.I, combined with our increased computational & processing power in the 21st century ([12] there has been a 1-trillion fold increase in processing power over the last 60 years) is excellent in processing large data sets and finding patterns within them, therefore we should be using them to get more accurate and smarter energy forecasts, i.e. forecasting supply, demand and price of the many forms of energy. How this would improve people’s lives is by guaranteeing a level of energy security ([13] defined as “the uninterrupted availability of energy sources at an affordable price”) by servicing the exact demand almost 100% of the time and not misusing supplies whilst also fundamentally tackling climate change. If we can reduce energy usage and production using this method of devising smarter energy forecasts, we can conserve fossil fuels and reduce the quantity being burnt, thus reducing an area’s carbon footprint (huge PMEs attached with this – SEE FIGURE 2). [14] In 2012, 9.7 billion tonnes of carbon were emitted into the atmosphere by burning fossil fuels alone – a testament to the threat it poses to the Earth.
The build-up of atmospheric carbon dioxide contributes towards climate change because the polarity of the compound means the bonds within it can absorb infra-red radiation emitted from the Earth, thus heating it up, and carbon soot causes global dimming by reflecting sunlight back into space and result in cooler than expected temperatures.
Not to mention the sustainment of the fossil fuels for future generations resulting from the employment A.I. in this way. Energy security is about long-term sustainability, not just about servicing today’s demand by exhausting supplies.
 |
Figure 3: ONE pathway of a PME from the reduced carbon footprint in an
area
|
The Education Sector
In the future, I think that schools and learning establishments will implement A.I. in the light of using it for personalised learning. [15] Every student has a different learning style, visual, aural, verbal etc, but not every style is catered for, leaving some students not being able to maximise their full potential which is not ideal in a world which is becoming dominated by tertiary and quaternary industries (essentially ones that are very knowledge, information and skill driven).
The A.I would provide and assist the teacher with information in multiple formats to improve the students experience and efficacy of learning. This same A.I can be reprogrammed into a ‘chat bot’ perhaps, and acting as a portable, personalized tutor accessible from any device. Multiple A.I chat bots have already been coded for recreational use e.g. [16] Cleverbot, which is constantly learning from its users. I think something similar, but in terms of providing educational, tailored resources and guidance to students of a certain learning style can be achieved.
For example, let’s say you have an aural learner, the A.I could instantaneously combine a series of pitches, rhythms and specific subject content to create a song, making the content more understandable, memorable and applicable.
A 24/7, 365 days a year support network consisting of A.I chat bots/tutors, which would be personalized to individual students? Seems quite far fetched, but the [17] Georgia Institute of Technology are already using an A.I. teaching assistant, named Jill Watson which has been reported to optimize learning and reduce the workload of Ph.D. students, whilst also adopting a colloquial tone to seem almost real. [18] Ashok Goel (the developer of Jill Watson) states that his own students did not realise that their teaching assistant was not human which is important in the sense that even with an A.I machine, human-like interaction can still be simulated.
 |
Figure 4: Snippet of a real
conversation with Jill Watson
|
This shows that the foundations are present, we just need to develop onto what we have and expand the usability to reach and affect as many students as possible. The questions answered usually have firm and objective solutions, but Ashok Goel sees this as the basis of a start-up in providing lucrative tools in the education field and
Not only would an A.I. of this sort cost an incredible sum ([19] building one chat bot can cost anywhere between $40,000 to $100,000 depending on complexity) but logistically would be hard to implement. [18] Ashok Goel had to collate four semesters worth of data, including over 70,000 questions and answers: only then could he start ‘training’ it and tailor it to his specific course – a very lengthy process indeed. Even then, Jill Watson was not an instant success, because initial testing from Ashok showed that she was giving incorrect information and even to this day she is only 97% accurate.
To make this somewhat of a wide scale system, strict trials should take place, showing off the benefits from using this system. Starting off with a few learning establishments – students within it can be divided into two groups, one using the A.I. alongside their teacher and one without - their performance can be monitored to draw conclusions. Even though they are not directly linked to knowledge, cognitive function tests devised by neuroscientists could also be used to show if A.I. based learning has any other advantages.
To Conclude:
[20] Louis Columbus, a contributor to Forbes claims that “Artificial Intelligence will enable 38% financial gains by 2035,” but if these ideas are implemented, I believe that it will direct A.I to be an improving force (in terms of quality of life), rather than merely a force for profit and financial gain.
The profit that I am referring to, which is not necessarily monetary includes something like acquiring a skilled, educated workforce who have nutritious diets to increase their productivity. This would be seen as tremendous ‘profit’ to a newly emerging economy or low-income country because it would ideally stimulate their economic development. The monetary profit would arise as the result of the improving force.
References:
[1] Rouse,
Margaret. 2010. "What Is AI (Artificial Intelligence)?". Searchenterpriseai. http://searchenterpriseai.techtarget.com/definition/AI-Artificial-Intelligence.
[2] Mayo, M. (2016). Basics of Data Science: What Patterns can be Mined?
Retrieved from
https://www.kdnuggets.com/2016/12/data-science-basics-types-patterns-mined-data.html
[3] Abadicio, M. (2019). Artificial Intelligence at
the FBI. Retrieved from
https://emerj.com/ai-sector-overviews/artificial-intelligence-fbi/
[4] S. (n.d.). Number of artificial Intelligence
(AI) Start-ups Worldwide in 2018. Retrieved from
https://www.statista.com/statistics/942657/global-ai-startups-by-country/
[6] Press, G. (2019). Indicators of the State of
AI. Retrieved from
https://www.forbes.com/sites/gilpress/2019/04/03/7-indicators-of-the-state-of-artificial-intelligence-ai-march-2019/#453f5a18435a
[8 W. (2016). Potential Economic Losses from
Uncontrolled Weeds. Retrieved from
http://wssa.net/2016/05/wssa-calculates-billions-in-potential-economic-losses-from-uncontrolled-weeds/
[9] McDivitt,
P. (2018). Does GMO Corn Increase Crop Yields? Retrieved from
https://geneticliteracyproject.org/2018/02/19/gmo-corns-yield-human-health-benefits-vindicated-21-years-studies/
[10]
Karpyn, Allison, and Sarah Treuhaft. 2010. "The Grocery Gap: Who Has
Access To Healthy Food And Why It Matters."– A
research paper by Policy Link & The Food Trust
[11] M. (2014). Inadequate Food Distribution
Systems. Retrieved from
http://12.000.scripts.mit.edu/mission2014/problems/inadequate-food-distribution-systems
[12] Routley, N. (2017). Visualizing the Trillion-Fold Increase in
Computing Power. Retrieved from
https://www.visualcapitalist.com/visualizing-trillion-fold-increase-computing-power/
[14] Levin, K. (2013). Carbon Dioxide
Emissions from Fossil Fuels. Retrieved from
https://www.wri.org/blog/2013/11/carbon-dioxide-emissions-fossil-fuels-and-cement-reach-highest-point-human-history
[17] IBM Cognitive Business. 2016."Would You Know
If One Of Your Teaching Assistants Was A Bot?". Medium. http://www.medium.com/cognitivebusiness/would-you-know-if-one-of-your-teaching-assistants-was-a-bot-5772cb8f7b68.
[18] Leopold, T. (2017). A Professor Built an AI TA
for his Courses - and it Could Shape the Future of Education. Retrieved from
https://www.businessinsider.com/a-professor-built-an-ai-teaching-assistant-for-his-courses-and-it-could-shape-the-future-of-education-2017-3?r=US&IR=T
[20] Columbus, Louis. 2017.
"Artificial Intelligence Will Enable 38 Profit Gains By 2035". Forbes. https://www.forbes.com/sites/louiscolumbus/2017/06/22/artificial-intelligence-will-enable-38-profit-gains-by-2035/#693d213d1969.
Figure References:
[1]: "Flying Quadcopter Drone For
Agriculture". 2015. Flickr. https://www.flickr.com/photos/ackab/15712761628
[3]: Created by myself using www.draw.io
[4]: Goel, Ashok, and Lalith Polepeddi.
2016. "Jill Watson: A Virtual Teaching Assistant For Online
Education."
very interesting
ReplyDeletehello, who is the fake above me? my one though is 'very interesting'
ReplyDelete