Proactive wildfire prevention Analytics

Today, wildfire risks are managed with reactionary risk-transfer model. The property owner (household/business) secures property insurance. In event of wildfire, the insurance company pays the property owner. The insurance company, in turn, manages its own risk via Reinsurance. Loss from each wildfire is huge. We believe proactive assessment and wildfire prevention/mitigation provides significantly improved reward. Proactive wildfire mitigation options, as opposed to reactionary ones, can significantly reduce risks and losses. Less expensive and more strategic risk/reward modeling with supporting data can change stakeholder business plans for proactive investment. Preliminary/isolated data points exist to support this claim. We plan to integrate these in a reusable model for stakeholder commitments and demonstrate feasibility in select properties. Scaled globally, the solution can mitigate wildfire in other parts of the world. The analytics/modeling will consider local ecological conditions toward mitigation recommendations. It positively enhances life, including: reduced pollution, improved health, ecological, carbon stock

We see the traditional wildfire risk-response model favors risk transfer, where involved parties continuously avoid responsibility for the wildfire. Recent studies, including Columbia University’s SUMA Net Impact, revealed a systematic problem in our current risk-response system. There is a dangerous bias towards addressing negative risk of managing financial loss/harm after wildfire. However, we also avoid the potential gains of positive risk (financial gains, or loss avoidance)obtained by preventing wildfire. This bias, combined with an underemphasis on prescriptive analytics and the undervaluing of passive risks, leads to potential losses.. These factors create deep distortion misdirecting stakeholders and preventing systemic change; and furthermore institutions neglect potential opportunities in their projects.

We firmly believe the situation needs remedy, regionally and globally. We see profound opportunities for institutions to leverage positive risks, create a more sustainable future and a more profitable institution. We confidently propose  social innovation as a solution to interrupt cycles of risk and accelerate a global shift from risk transfer to risk prevention. Our role is to provide the methods and analytical models to insitigate change.

We expect in short term the adoption will be local/project-specific; however expanding to global once stakeholders see tangible benefits.

We have an innovative analytics-based approach to wildfire prevention/mitigation.  We plan to use data analytics that takes input

• Historical data from various fire conditions
• affected land size and population
• ecological condition, topography, causes of fire, population and land affected
• layers of fire prevention measures: such as herd grazing to reduce flammable material, fire resistant plantation, house hardening, etc
• financial losses incurred by various stakeholders (e.g., insurance and reinsurance)
• Comparable information on selected area for wildfire prevention measures to be implemented
• Financial projections
• Implementation cost of various prevention measures
• Reduction in loss (aka savings, positive risks) from the wildfire prevention

The analytical model will generate

• Recommended and optimized different levels of wildfire prevention 
• An estimation of costs for each stakeholder segment
• Savings (cost avoidance) for each stakeholder segment
• Total cost/benefit summary
• Potential contribution for each stakeholder segment  to implement wildfire prevention steps

We expect

• Participating stakeholders for affected region will form a consortium to manage/monitor/track contributions and implementation costs.
• CrowdDoing will continue to enhance the modeling and tools for evolving region-specific conditions; and support stakeholders (contributing and affected) with assessments and recommendations.
• CrowdDoing delivering valuable service for net risk reduction

Abstract

The 2030 Agenda for Sustainable Development adopted by all United Nations Member States in 2015 provides a shared blueprint for peace and prosperity for people and the planet, now and into the future.  At its heart are 17 Sustainable Development Goals (SDGs).  These broad areas represent the earth’s most urgent challenges. CrowdDoing, an organization that uses skilled volunteers to tackle global challenges, has created a wildfire prevention modeling strategy in response to four SDG’s:  Good Health and Well-Being (SDG 3), Clean Water and Sanitation (SDG 6), Climate Action (SDG 13) and Life on Land (SDG 15).

Leveraging artificial intelligence, machine learning, and data science, CrowdDoing has developed a functional model called Prevention Derivatives, which integrates positive and negative risks to prevent wildfire.

Traditional risk management stakes its relevance in anticipating and responding to a given event’s possible negative effects. This close focus tends to sacrifice those positive outcomes which coexist with negative risk and results in institutions compensating for previously unknown risks merely as the “cost of doing business.” The Prevention Derivatives model validates the concept that assessing and preventing anticipated risks is cost-efficient, even to the degree that it can monetize the capitalization of positive risks. We hope to find fellow innovators hoping to lead a global movement toward risk prevention. 

The World’s Positive Risk Blindspot   

Prevention is better than cure; and with significant financial benefits. Healthcare industry allow free flu-shots to improve community health and reduce overall costs. Auto industry mandates safety-belt for improved public safety and reduced life/health insurance costs. And, yet, no such program exists for wildfire prevention.

At CrowdDoing, we believe that we can address the persistent threat to sustainability that our planet faces through systemic change. Due to underfunding and a lack of institutional attention to sustainability goals, global sustainability faces a crisis.  A principal factor in this equation is the current Risk Management approach adopted muti-institutionally. The 2020 global pandemic has engendered devastating effects for the economic, political, and environmental future. As a result, we are forced to rethink and face threats that have been avoided and transferred for years. Rapidly increasing globalization makes restructuring how we approach risk more urgent. With increased connectivity, a once-isolated risk has the potential to affect the increased potential for another. While significant trade increases have had many positive effects on economies, it brings globalization of health risks.  Coupled with the effects of climate change raising the consequences of individual risks, there appears a pressing demand for large-scale change to our approach to risk.

Currently, the traditional risk-response model favors risk transfer, wherein various parties avoid responsibility for the impending crisis. This model is becoming increasingly fatal. Recent studies, including Columbia University’s SUMA Net Impact, have revealed a systematic problem in our current risk-response system. A dangerous bias towards addressing negative risk, meaning the possible financial harm if a risk occurs. However, in doing so, we also avoid the potential gains of positive risk, which are the benefits obtained by preventing that risk. This bias is combined with an underemphasis on prescriptive analytics and the underpricing of passive risks, which are potential losses due to inaction. These factors have come together to the market's detriment, creating deep distortion that misdirects stakeholders and prevents systemic change. Institutions neglect potential opportunities in their projects due to the deep distortion in the market.

We firmly believe that this situation can be remedied. We see profound opportunities for institutions and companies to leverage positive risks to create a more sustainable future and a more profitable institution. While general risk management offices attempt to address obvious risks, they are not given the latitude to control budgets, provide risk metrics, and anticipate opportunities. At CrowdDoing, we firmly believe in the power of social innovation to interrupt cycles of risk and accelerate a global shift from risk transfer to risk prevention. Our role is to provide the methods and models that will begin this change.

A Case for Prevention vs. Response

We are proposing a risk prevention strategy to complement and challenge the risk response strategy currently in use. Risk prevention will reduce the cost of risk transfer for companies over the years and help meet our sustainable development goals. As a part of this new strategy, we will use a portion of the cost reduction to implement scientifically proven solutions that prevent significant sustainability threats, including disease prevention, poverty, and natural disasters.

CrowdDoing focuses on transforming the insurance industry from a bias towards risk transfer to a more forward-looking approach taking positive risks that increase returns associated with prevention measures. This approach allows us to finance “upstream parametric interventions.” To fully address this concept, it is essential to understand how parametric and upstream interventions are merging to facilitate prevention innovations. Parametrics, put broadly, are the tangible measurements that trigger financial intervention. For example, parametric through a downstream operation will define a metric that must be reached for claims to be distributed. Upstream parametric create standards that lower risk before the catastrophe can even occur, rather than compensating for its impact. For example, upstream parametric might dictate the fuel load on the landscape proportionate to the risk that it brings within a geometric distance.

Put simply, we have found that it is more economically and environmentally effective to prevent risk before it occurs rather than gamble on its probability. Through data science, statistical models, and archives in other domains, we compare the cost efficiency of risk prevention to the cost efficiency of risk transfer for multiple stakeholders. Multiple industries, such as healthcare and auto insurance, have demonstrated that effective prevention is more cost-effective than post-loss response. As an analogy, two decades ago, the healthcare industry learned it was more cost-effective to offer flu shots for no cost rather than admit and treat an insured person in a hospital. Likewise, the car insurance industry earned higher profits when seat belt laws were in place, rather than compensating for loss after accidents at times without such laws.  “IIHS-The Insurance Institute for Highway Safety” campaigns for regulatory changes by the government on vehicle safety standards as a form of preventing risks.  Insurers have already assessed that market demand for prevention exists strongly among those who are insured: “ more than 55% of these customers are prepared to pay extra for risk prevention services,” according to World Insurance Report 2019. There are times in California where air quality is so poor that medical advice cautions residents not to leave the house. We believe there would be value in a system preventing these conditions from occurring, but that is not something that today’s insurance coverage can address. Therefore, CrowdDoing is looking for preventative contributions from California’s residents and institutions.

CrowdDoing focuses on transforming the insurance industry from a bias towards risk transfer to a more forward-looking approach taking positive risks that increase financial returns associated with prevention measures. This approach allows us to finance “upstream parametric interventions.” To fully address this concept, it is essential to understand how parametric and upstream interventions are merging to facilitate prevention innovations. Parametrics put broadly, are the tangible measurements that trigger financial intervention. For example, parametrics through a downstream operation will define a metric that must be reached for claims to be distributed. Upstream parametrics create standards that lower risk before the catastrophe can even occur, rather than compensating for its impact. For example, upstream parametrics might dictate the fuel load on the landscape proportionate to the risk that it brings within a geometric distance.

We believe it is worth emphasizing again the major fault in the current risk-transfer economy. The financial cost of not addressing risk (passive risk) is consistently underpriced, and combined with the bias against positive risk, market distortion occurs. Liabilities are treated either as inherent costs to doing business or un-predictable risks, even when entirely preventable. In response, administration devoted to risk management tends to avoid dangerous risks, neglecting abundant profit opportunities.

The CrowdDoing Alternative: Prevention Derivatives and Sustainable Change

Prevention Derivatives seeks to use a predictive machine learning model to estimate potential savings for a given geographical region. These savings would include stakeholders’ properties, business profits, everyday health, and regional ecology resulting from applying risk prevention solutions versus inaction.

We want to invert a company's natural hedges to benefit both the company and the landscapes at risk. Traditionally, natural hedges are the markets in which corporations and individuals will invest to mitigate loss in an associated market.  When a liability already exists in a landscape, you can reduce the risk and diversify against it through the intervention of targeted prevention derivatives. Our role in implementing this model is to examine an institution's liabilities and assess how preventable they are and provide a comprehensive financial summary. Put simply, our summary might present something like this: “for every dollar invested in prevention, you could save five dollars from anticipated liabilities. For those that are cost-efficient to prevent, we will provide necessary model and financial analysis that if this institution invests their money into scientifically proven solutions, they will say x$ on fire liabilities.  

Prevention Derivatives for Wildfire Prevention

We have identified the mitigation of wildfires and wildfire-associated risks as one of the most effective uses of Prevention Derivatives. We have dedicated extensive research in data analytics, statistics, and financial modeling towards finding a prevention derivatives solution to wildfires’ increasing challenges during a particularly fire-ridden decade. A spectrum of forest fire risk prevention approaches is available: from creative new social innovations to well-established interventions. One of the solutions  we have found is surprising, cost-effective, and has proven successful. Goats and sheep benefit from natural grazing in areas abound with flammable debris, grass, and thicket. We believe that prescribed, systematic goat and sheep grazing can provide cost-effective brush control, which will ultimately relieve corporate liabilities and make a positive impact on sustainability and landscape preservation issues. Finally,  an environmental sustainability benefit for the wildland-urban interface allows goats and sheep that would otherwise be confined and fed in ways that contribute to destructive animal and carbon emissions. When allowed to graze, we reduce animal emissions, fuel load, and ultimately, wildfire risk.

In any region, forest fire risk prevention modeling would first survey existing forest fire prevention interventions in the area with an eye towards which ones might be suitable to support and expand. We would then review global social innovations relevant to that region to determine which to replicate and co-invent social innovations where none are appropriate to reproduce. Our approach will identify comparison regions to assess relative risk prevention controlling for temperature, humidity, wind intensity, tree types, etc.  Using predictive data models, we can engage stakeholders for preventative solutions that will result in economic and environmental benefits.

Ultimately, we predict that prevention derivatives can result in the following:

• Lower short-term expenses directly related to wildfires (e.g. Suppression, Evacuation, property loss, etc.)
• Reduction in social, ecological, and health consequences
• Lower long-term expenses (Landscape Rehabilitation, Business Interruption, natural resource loss, etc.)
• Ecological control via regeneration of natural resources
• Continued adjustment to tools via monitoring and control steps for sustained viability against changing climatic condition

We use data analytics that can model and predict biodiverse and regenerative wildfire prevention measures suitable for a region. It uses cost models to predict institutional investments towards various preventive measures ranging from simple preventive grazing to wildfire protective landscaping to advanced robots creating fireproof perimeter; forecasts savings from wildfire avoidance, and makes recommendations for optimized risk/rewards. We expect stakeholders will use our model as guidance to implement prevention measures, share equitably the costs, each gaining financially. 

Stakeholders such as insurance/reinsurance companies need an established model. Our goal is to productize our concept, make it a thriving industry available globally.

We have established the concept that proactive measures for wildfire prevention provides significantly improved risk/reward benefits over cost from losses in the event of wildfire. Some stakeholders have accepted the concept in principle. However, for active participation, stakeholders are looking for a verifiable model.

We consider our activities in prototype phase since

• this work program will convert the concept to a risk/reward model,
• demonstrate model effectiveness in 2-3 select regions,
• identify model enhancements needed for these select test cases,
• work with stakeholders towards a wider acceptance of the risk-reward model, and formalize it amongst other institutions
• identify enhancements broader applicability. Time permitting, and availability of required data (ecological and financial), we will extend tests a few more properties to increase our confidence with the model, its extendibility, and future implementation.

Time permitting, we will extend tests a few more properties to increase our confidence with the model, and future implementation.

Our solution is disruptive – it forces current long-utilized risk-transfer model to be replaced by a risk prevention model. It requires institutions to work together to prevent wildfire in a community.

We see the traditional risk-response model favors risk transfer, where involved parties avoid responsibility for the wildfire.We firmly believe the situation needs remedy, regionally and globally. We see institutions have a large opportunity to invest proactively in wildfire prevention measures, create a more sustainable future and a more profitable institution. We see the preventive investments benefits community/society too.

Our solution is innovative in a theoretical sense, as we are breaking away from the traditional risk-response model. However, we are also using data analytics in a completely new way to make prevention recommendations, overall and objective cost/benefit summary, and recommended contribution of each stakeholder consistent with their financial gains. It allows UN ESG guidance to be implemented while making such implementation financially viable. It creates new business to implement the recommendations.

Our solutions target insurance/reinsurance and select large corporations also supporting both the innovative and disruptive nature of our solution.

As of now the solution is in prototype phase.

In one year we expect model will be ready to trial in 1-5 regions in California region. This will

• Validate the AI Model under implementation as well as allow us to make the enhancements needed
• Prove the effectiveness of our  recommended wildfire prevention methods and effectiveness to these properties
• Costs of implementation
• Cost contributions, expected savings, and ROI
• Alignment with stakeholders on model viability and its implementation

The properties will be selected jointly with insurance and reinsurance companies before broader rollout. Beneficiaries beyond insurance/reinsurance companies include all members in communities where wildfire risks have been reduced.

Over five years we expect the model can be easily adapted for new regions. With this

• The model can be used for other regions within the US
• The model can also be selectively used for regions outside the US (conditions closely aligning with US conditions)

The prevention derivative is focused on the following UN SDG areas

• Goal 3: Good Health and Well-being
• Goal 6: Clean Water and Sanitation
• Goal 9: Build resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation
• Goal 13: Take urgent action to combat climate change and its impacts     
• Goal 15: Protect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity 

Mapping of our activities to SDG initiatives can be found in link https://drive.google.com/drive... .

Implementing wildfire prevention measures leads to

• Sustaining land and forest
• Sustaining community and local ecology
• Avoiding pollution, including toxic fumes, from wildfire; improved community health
• Avoiding airborne pollution to far-away communities
• Avoid further pollution of groundwater from seeping toxic water
• And best of all, with proactive investment from insurance/reinsurance companies while they also gain positive financial returns.

Success for the initiative will be measured with

• The amount of stakeholder participation in multi-party agreement for proactive investment in wildfire prevention. In addition,
• the net risk reduction of wildfire damages and wildfire air pollution

Currently all CrowdDoing activities are supported with volunteer resources. We have about 50 active part-time volunteers supporting wildfire prevention derivatives. In addition, we also have part-time University pro-bono team contributing to the initiatives.

CrowdDoing personnel are expert in their respective fields and currently involved in cutting-edge research and development of wildfire risk prevention strategies. Key personnel and their skills include:

• Bobby Fishkin: Lead, Co-Founder
• Darius Alexander: Architecture
• Goutam Sinha: Chief Business Officer
• Mikhail Teverovskiy: Data Science Manager
• Tom Watson: Lead Designer
• Wendy Nystrom: Chief Risk Officer
• Tobias Temmen: expertise in banking and insurance industry

Additional information on personnel can be found in https://drive.google.com/file/d/1Cph1wepO-c3lcMpAhV4YTiAMB3orbqmB/view?usp=sharing .

CrowdDoing non-discriminatory policies ( please see https://drive.google.com/file/d/176IeYZtXJ4dBAKIAnvTG0N2MVbZxQzYx/view?usp=sharing ) outlines our principles to build a diverse, equitable, and inclusive team. Our global volunteer base and partnership with diverse pro-bono universities further demonstrate our continued emphasis on these principles. 

The solution to our current wildfire crisis is driven by data analytics and AI. By nature, we  need a large volume of historical data, constant analysis of new information, and model adjustments. Due to nature of data/analysis in an early stage and complex market – we now require a key group of dedicated resources that can be difficult with solely volunteer resources. We see Solve funding as a way to secure a team of employees towards the project.   I see the need of a key group of dedicated resources that is difficult to achieve with pool of volunteer resources. I see Solve money as a way to overcome the condition to move the project to fast track.

Primary area where we need support is in area of business modeling. 

As noted in the reference letter https://drive.google.com/file/d/18GHAblBA3k3IEQ5L73_brOe3H5MH_OFI/view?usp=sharing – the risk reduction principles have already adopted by insurance industries such as healthcare and auto. However, the concepts are new to property insurance/reinsurance segments.

The success of the change model depends on a business model well understood by insurance/reinsurance industry. The model

• Needs to be consistent with insurance industry terminology and processes
• Demonstrate how the risks are managed
• How it is financially beneficial to the company, and industry, without causing incremental financial strain to property owners
• How it maps to their ESG objectives
• How AI and analytics based approach provides them a market advantage
• How in the long term the model can be converted to a financial market derivative

The AI and analytics model can also be used by Fortune-500 companies to reduce their own liabilities for better financial returns (e.g., for companies with self-financed insurance). Beyond the model considerations above, they will also be interested how the model can be extended to support their carbon footprint commitments made to market and investors.

We see MIT and MIT partner knowledge of market, industry trend, and research knowledge can help us significantly in implementing the socially innovative change model.

We are looking for MIT team support in the following areas

• Common data/database sources of fire affected regions that CrowdDoing can access. It is best of the access is free of charge. This is needed to build and to test the model
• A sounding board for validation of our model and approach
• Partnership in papers/presentations extending visibility of social innovation to others
• Collaborating with environmental science team
• Collaborating with insurance agency contacts
• Access to simulation tools related to wildfire, where available

Our project is on social innovation to reduce wildfire; to create smart, safe, and sustainable communities around the world. Once prototyped and implemented, we see the solution is applicable to global community. Thus we see it is aligned with GM’ Resilient Ecosystems Challenges.

We engage volunteers globally in such initiatives.

Our solution uses data science, analytics/AI/ML, historical cost data from past wildfire events (i.e., cost of no action), environmental conditions for a new select region, cost of implementing wildfire reduction steps and generates recommendation for a financially beneficial wildfire reduction plan.

Benefits we see beyond avoidance of property loss include (examples):

-             Improved health (less needs to visit doctor from asthma and others due to reduced air pollution)

-             Improved participation at work (fewer air pollution days)

-             Create and support alternate new low-cost business in the community (e.g., goat grazing, fire retardant gardening)

With above we would like to be considered for The AI for Humanity Prize.