Ai Predicts California’s Greatest Challenges of the next 5 years

by Peter Koehler


Posted on 7/22, 2018 at 2:22 PM


bay area technical sales

Artificial Intelligence is continuously being adopted to improve processes and solve problems in technology, medicine, and industrial applications, but often AI is used to help make predictions. Common Ai models are used to predict consumer behavior, market trends, and numerous other areas of business and social logistics. In a recent project we explore a variety of Ai models to look into how Ai predictions could be used to look at some of the greater challenges that we face, as well as what we are doing to help identify and solve some of these challenges. The complexity of such a project should be noted and we took into careful consideration the age and dynamics of the data, the sources of data, and the popularity of the data. Initially a hand selected large variety of data on many issues, including a variety we had limited knowledge of prior to this project was used to help measure response values and weights for the top challenges as they began to emerge. Although not every challenge could be assessed and not every case be collected a good portion of data was collected by interpreting controversial issues and topics that currently impact or have potential to effect larger portions of the population. Certain challenges hold more value and are weighed based on the potential impacts they represent, with models being developed to give higher weight values and priority to those issues that effect the most people.

 It doesn’t take an AI model to tell us that these challenges exist, or that health care, pollution, economic disparity, immigration, gun control and plenty of other challenges are all things we face today and will likely face over the next 5 years. What it can show us is an outside perspective, an AI viewpoint of how some of these challenges compare to others and how this machine perspective sees which challenges are of most concern. An added benefit of seeing the challenges throughout the project was being able to see the various pieces of each challenge along the way. The model identifies the complexity of some of the problems and adds a new ability to see these challenges through various independent metrics that could be isolated and easier to interpret if action was to be taken to work with these challenges in the future.

A majority of the data comes from commentary on social media and uses various methods of interpretation including sentiment analysis and response activity to measure the results of numerous existing and specifically generated posts. Other data includes a single proprietary project where data was collected in surveys and through the research of material used in a personal college paper. Additional data and information was gathered prior to the models inception that included a selection of different demographics and sources. The results are only hypothetical and can be seen as only a small snapshot of what may be some of the more difficult challenges we are likely to face in the near future. The effect of media coverage or specific events around any single challenge can great amplify and skew the data collected around these spikes, and measures were taken to balance the data collected during such events. The results below are listed in no particular order and were chosen for inclusion based on the number of people these issues were likely to impact. Though the Ai models are not able to necessarily present us with challenges we are not already aware of, it was able to help present new perspectives into the various pieces that make up these challenges

Fake News - Fictional Vs Factual and Real information.  As fast as information moves today, it is now one of our greatest challenges to determine if the information we receive is real or is created for the benefit of some other purpose. Fake news can be just as convincing as the real thing and to the average person it may be nearly impossible to determine when the media is fake or altered. Fake information can be used to influence markets, develop social movements, skew opinions, and sell us products out of fear, hype, and other methods, based entirely on fictional information. With software, almost anyone can create fake news with fake audio to make someone say something they didn’t, and even create fake video as false proof that someone did something they did not. This is a challenge that we must consider as highly dangerous and complex as media studios, corporations, and other groups can create fake news, fake evidence, and other fake information that has nothing to do with reality. The ability to influence people, generate distractions, and execute objectives with information that is not true is being used more and more on a daily basis. 

Immigration - A hot topic these days with both sides holding strong to a complex issue that seems to have no clear cut solution and will definitely occupy numerous resources to resolve in the near future if that is even possible? Testing an AI models ability to go beyond logical processes and include empathy, conditional interpretations and other factors in this case proved to be a challenge in itself.

Disaster - The chances of experiencing a major natural disaster or man-made disaster in the near future appeared to be a common assumption in this data pool. Perhaps this was just by pessimistic chance, but obviously a major disaster will also pose a major challenge if and when it does occur.

Health Care - Although this is a broad category we chose to use, we grouped health care with a number of other related social sub categories and factors that contribute to this complicated challenge. Health Care costs, complicated legal and business practices, government and private policies are all entangled in a system that would not be easy to un-wind. The only real solution maybe to start over. The promise of Ai models to help find solutions in this case looks promising, but getting the people involved to agree, and the resources to make the changes happen will surely be a much greater challenge.

California Water Crisis - Water pollution, availability, and water rights are all challenges that need solving if we are going to continue to have access to clean water. As many nations improve the access and technologies around water quality and availability we may be heading in the opposite direction. The data around this particular challenge was limited and more research is warranted to further analyze and validate the information. The water challenges in California could affect millions of people, and most of the states fresh water. We will be faced with the challenges of water rights, water pollution from pesticides, drugs, and other chemicals, that are currently finding their way into our water supply. Many of these new contaminants are never tested for and not being screened for in current water quality test and evaluations. Many people, may know very little about the challenges around our water systems or the legal and political complexities involved. Was a large portion of California’s water privatized and sold in a highly complex combination of legislation in the 1990’s. ( http://www.kwb.org/store/files/110.pdf) Recent proposals to further privatize water rights could lead to distribution limits, higher costs, and numerous legal battles. Natural droughts and other challenges surround our water and the responses and data provided for this challenge definitely warrant some attention.

State of Jefferson - ok, we probably said enough, its time to go, thanks for reading……

These challenges are all complex and since they will involve so many people hopefully we can use our own thinking and Ai to help us better understand and solve them or work on them together. Perhaps it is just a choice we should consider to look at each of the challenges we face as a society and use Ai models to help expand our awareness into what we can do to improve.

Peter Koehler is a contributor of technical articles and documentation related to cognitive computing and artificial intelligence. [email protected]