Investing In AI. Where To Start? Lost? Don't Get Left Behind.

A Clear Path To Profitable AI Investing (RAIR) & (IRAI).

Sep 10, 2024

Welcome to Silicon Sands News, read across all 50 states and 96 countries. We're Silicon Sands Studio and 1Infinity Ventures, and we’re excited to present our latest edition on how responsible investment shapes AI's future, emphasizing the OECD AI Principles. We're not just investing in companies—we're investing in a vision where AI technologies are developed and deployed responsibly and ethically, benefiting all of humanity.

Our mission goes beyond mere profit— we are committed to changing the world through ethical innovation and strategic investments.
We're diving deep into a topic reshaping the landscape of technology and investment: Today, we will explore a topic demonstrating how responsible investment in AI can positively impact the global society—that topic is defining industry metrics for venture capitalists to measure responsible investments in AI based on the OECD AI principles. This is not just about numbers on a spreadsheet; it's about creating a framework that ensures the AI we develop today will lead to a better tomorrow.

TL;DR

This article explores the critical and urgent need for responsible AI investment and proposes a comprehensive framework based on the OECD AI Principles. It introduces the basis for quantifiable metrics for each principle, designed to evaluate AI startups and technologies. The piece highlights two key aspects of responsible AI investment: funding tools that enable ethical AI development and supporting startups committed to responsible AI practices. It emphasizes the importance of clear communication between startups and investors regarding ethical AI practices. The article outlines the formation of a Working Group on Investment in Responsible AI (IRAI) to refine and implement these metrics. It proposes the creation of a Responsible AI Investment Rating (RAIR) as an industry benchmark and the creation of an independent, industry-supported NGO to maintain the benchmark. Challenges in implementing this framework are addressed, including the need for standardization and the balance between ethical considerations and innovation. The article concludes with a call to action for the entire venture capital ecosystem, urging collective adoption of these principles to drive responsible AI development and shape a more ethical, sustainable AI future.

Two Aspects of Responsible AI Investment

Some might ask why they should invest in responsible AI. I would ask, how do you expect AI to be successfully implemented if it needs to be developed responsibly? It is human nature to avoid things they don’t trust, and responsibility is essential to trust. On top of this, there is mounting consumer demand for responsible AI practices and growing governmental oversight in the form of regulation. With this in mind, responsible AI is not just investable—it should be a requirement for any AI investment.

There are two distinct investment opportunities in AI: Tools that enable responsible AI and AI that is developed responsibly. On one hand, there’s a growing market for tools and technologies that facilitate responsible AI development. On the other hand, there’s a critical need to invest in AI startups that adhere to responsible development practices. This dual approach broadens the scope of potential investments and reinforces the ecosystem of ethical AI.

The first aspect involves identifying and backing startups and creating the infrastructure for responsible AI. These could be companies developing advanced bias detection algorithms, making more energy-efficient machine learning models, or building robust privacy-preserving data analysis tools. VCs can help lay the groundwork for a more responsible AI ecosystem by investing in these enabling technologies. These tools become the building blocks other AI startups can use to ensure their developments align with ethical principles.

Simultaneously, VCs must also focus on investing in AI startups that demonstrate a commitment to responsible development of their products and services. This means looking beyond an AI solution’s technical capabilities or market potential. It requires a deeper evaluation of the startup's processes, values, and long-term vision. Are they considering potential biases in their training data? Have they thought about the environmental impact of their models? Do they have clear policies on data privacy and user consent? These are the questions that responsible AI investors must now consider alongside traditional metrics.

However, this dual focus introduces a new challenge: how can startups effectively communicate their commitment to responsible AI to investors and customers? As AI ethics evolves, there's a growing need for a common language and framework that allows startups to demonstrate their responsible practices easily.

This is where our proposed benchmark and metrics come into play. By providing a standardized way to measure and report on responsible AI practices, we can create a common ground for communication. Startups could use these metrics to articulate their ethical standings, much like they currently use technical specifications or market size estimates to convey their potential.

Imagine a future where a startup's pitch deck includes its total addressable market and projected revenue and its Bias Mitigation Score or Environmental Sustainability Index. Or where a company’s website features its Ethical AI Certification alongside its product features. This level of transparency would aid VCs in their investment decisions and help customers make informed choices about the AI products and services they use.

This transparent communication of responsible AI practices could become a significant competitive advantage. As public awareness of AI ethics grows, startups demonstrating their commitment to responsible development may find it easier to attract investment and customers. It could differentiate them in a crowded market and build trust with users increasingly concerned about the ethical implications of the technology they interact with daily.

The challenge lies in making this communication straightforward and accessible. It shouldn't require a PhD in ethics for a startup founder to articulate their responsible AI practices, nor should it demand extensive resources that might disadvantage smaller players. This is where industry-wide standards and easily implementable tools become crucial.

By investing in both the tools for responsible AI and the startups practicing responsible AI development, VCs can help create a virtuous cycle. The more we invest in enabling technologies, the easier it is for startups to implement responsible practices. And the more startups demonstrate the value of responsible AI, the greater the demand for these enabling tools.

This dual approach to responsible AI investment, coupled with clear and accessible communication standards, can transform the AI landscape. It can help ensure that as AI continues to permeate various aspects of our lives, it does so in a way that is ethical, sustainable, and aligned with human values. For VCs, this represents not just an opportunity for financial returns but a chance to shape the future of technology in a profoundly positive way.

The Foundation

Before digging into the metrics and benchmarks for responsible AI investment, we need to understand the foundation for these ideas. The Organisation for Economic Co-operation and Development (OECD) AI Principles, adopted in May 2019, provide a comprehensive framework for developing trustworthy AI systems. These principles have been embraced by OECD member countries and beyond, representing about 50% of the global GDP and serving as a worldwide reference point for ethical AI development.

The OECD AI Principles comprise five fundamental values:

AI should drive inclusive growth, sustainable development, and well-being. This principle emphasizes that AI systems should be designed to improve the lives of individuals and advance positive outcomes for society. It recognizes AI's potential to address global challenges and improve quality of life while ensuring that the benefits of AI are broadly shared across society.

AI systems should respect human-centered values and fairness. This principle underscores the importance of AI systems that respect human rights, democratic values, and diversity. It calls for AI designed to promote fairness, avoid unfair bias, and ensure equality of opportunity.

The third principle focuses on transparency and explainability. It advocates for transparent AI systems whose human users can understand decision-making processes. This principle is crucial for building trust in AI systems and ensuring accountability.

AI systems should be robust, safe, and secure. This principle emphasizes the need for AI systems that function reliably and safely throughout their lifecycle. It calls for potential risks to be continually assessed and managed and for AI systems to be resilient against attacks and misuse.

AI systems should be accountable. This principle states that organizations and individuals developing, deploying, or operating AI systems should be held accountable for their proper functioning per the above principles.

These OECD AI Principles provide a comprehensive ethical framework for AI development. They recognize AI’s transformative potential while acknowledging the need to address its risks and challenges. By grounding our approach to responsible AI investment in these principles, we ensure that our efforts align with globally recognized standards and contribute to the development of AI that is innovative but also ethical, safe, and beneficial to society.

As we discuss metrics and benchmarks for responsible AI investment, these principles will serve as our north star, guiding our approach to evaluating and fostering AI technologies that are profitable, trustworthy, and aligned with human values.

A Metric for Every Principle

The OECD AI Principles have been adopted by countries representing about 50% of global GDP. In addition, most frameworks, standards, and regulations are based on these principles. They are widely known and adopted by governments, NGOs and corporations alike. For these reasons alone, it makes sense for the investment community to adopt them as the basis for a metric system to hold ourselves and our portfolio companies accountable.

At 1Infinity Ventures, two partners have been involved in the responsible AI community since its inception. From this, they have a deep network of experts to tap into as we develop this framework: technical experts from all over the world in AI, policy, legal, ethics, humanities and other domains.

This article extends a previous reference to the investment community developing a quantitative metric system for AI investments. We developed these metrics to provide a quantitative framework for evaluating AI investments, ensuring that we're not just paying lip service to responsibility but actively measuring and promoting it.

Our proposed set of Responsible AI Investment Rating (RAIR) metrics are as follows:

1. Inclusive Growth, Sustainable Development, and Well-Being

Social Impact Score (SIS): This metric quantifies an AI system's impact on underserved communities as measured by the inclusivity of underserved communities
Environmental Sustainability Index (ESI): This metric measures the percentage of an AI system's environmental footprint using verifiable green energy.
Job Transformation Rate (JTR): This metric measures the net effect on employment as measured by human replacement versus human augmentation rate.

2. Human-centered Values and Fairness

Bias Mitigation Score (BMS): This metric assesses an AI system's ability to detect and mitigate biases. We aim to reduce biased outcomes, ensuring that AI systems don’t perpetuate or exacerbate societal inequalities.
Diversity of Training Data (DTD): This metric ensures that AI systems are trained on data representing the community the AI system will impact.
User Empowerment Index (UEI): This measures how much control and understanding users have over the AI systems they interact with. It is measured by the ease with which a human can opt out of using the AI system without penalty.

3. Transparency and Explainability

Explainability Quotient (EQ): This metric quantifies how well an AI system can explain its decisions in human-understandable terms.
Algorithmic Transparency Score (ATS): This score measures an AI company's openness to its algorithms and data usage, as measured by the publication of research papers, open sourcing of non-critical components, and clarity in data privacy policies.
User Feedback Loop Efficiency (UFLF): Tracks how effectively and quickly user feedback about the clarity and accuracy of explanations is incorporated into the system as measured by the time taken to resolve user feedback and update explanations for misclassified or unclear outputs, measured in hours or days.

4. Robustness, Security, and Safety

Adversarial Resistance Index (ARI): This measures the percentage of AI system outputs that are incorrect or maliciously altered when subjected to adversarial inputs as measured by the Ratio of incorrect outputs to total outputs in response to adversarial test cases.
Security Vulnerability Detection Rate (SVDR): tracks the system’s ability to detect security vulnerabilities, such as unauthorized access attempts, injection attacks, or other breaches, as measured by the Percentage of successful detections versus total security breach attempts.
Incident-Free Operating Time (IFOT): this measures the amount of time the system operates without causing harmful incidents (such as generating unsafe outputs or facilitating dangerous actions) as measured by the Average time (in hours or days) between safety-related incidents or system failures.

5. Accountability

Ethical Governance Score (EGS): This metric evaluates the strength of a company's AI ethics governance structures. It includes factors like an ethics board, clear escalation procedures for ethical concerns, and integration of ethical considerations in the development process.
Individual Accountability Metric (IAM): This assesses the degree to which individual roles within the organization are clearly defined concerning AI ethics and decision-making. It considers factors such as designated ethics officers, ethics training for AI developers, and personal performance metrics tied to ethical AI development.
Incident Response Efficiency (IRE): This assesses how quickly and effectively a company responds to ethical, legal, or operational issues with its AI systems. It could be measured in terms of response time, resolution effectiveness, and transparency in communication. It also evaluates the clear assignment of individual responsibilities in the incident response process.

The Investor Imperative: Catalyzing Responsible AI

As Investor, we're not passive observers in the AI revolution. We're the fuel that powers the engine of innovation. We can drive the entire AI ecosystem towards more responsible development by adopting these metrics in our investment decisions.

Imagine a world where every AI startup pitches its technology, market potential, bias Mitigation Score, and Environmental Sustainability Index. A world where Investors, Governments and Institutions compete not just on financial returns but also on the positive impact of their AI portfolios.

The reality is that the decisions we make today as Investors will shape the AI landscape of tomorrow. By adopting these metrics, we are setting a revolutionary new standard for AI investment, and creating a universe of opportunities. Companies that align with these principles are more likely to build sustainable, trustworthy products that stand the test of time and regulatory scrutiny.

This isn't just altruism—it's a forward-thinking business strategy. As public awareness of AI ethics grows, companies prioritizing responsible development will have a significant competitive advantage. They'll be better positioned to navigate an increasingly complex regulatory landscape and win the trust of consumers and enterprise clients alike.

Consumers are demanding transparency and fairness from the technologies they interact with daily. By investing in companies that meet these demands, we're not just doing good but positioning ourselves at the forefront of a significant market shift.

Challenges and the Road Ahead

We need industry-wide standards. This will necessitate unprecedented collaboration among venture capitalists, AI companies, ethicists, and regulators. We're fostering a new ecosystem of responsible innovation.

The complexity of measurement presents another significant hurdle. Quantifying abstract concepts like fairness or transparency is inherently challenging. Our methodologies will need continuous refinement as we gain new insights and face new ethical dilemmas. This is not a one-time solution but an ongoing process of learning and adaptation.

We must also grapple with the tension between short-term and long-term considerations. Optimizing for these ethical metrics may slow development or increase costs in the short term. Lets shift the narrative from seeing ethics as a constraint to recognizing it as a catalyst for sustainable growth and innovation.

Perhaps most critically, lets not make these metrics becoming mere box-ticking exercises. Our goal is not to create a new form of compliance theater but to drive genuine commitment to ethical AI development. This requires a cultural shift within the AI and investment communities, placing ethical considerations at the heart of the innovation process rather than treating them as an afterthought.

Despite these challenges, the potential rewards of this endeavor are immense. By pioneering these metrics, we can shape an AI future that's not just technologically advanced but also ethically sound and socially beneficial. We stand at the threshold of a new era in AI development, where ethical considerations are not constraints but competitive advantages. By embracing this challenge, we can help ensure that the AI revolution enhances human potential, bridges societal divides, and addresses our most pressing global challenges. The road ahead may be complex, but the destination - a future where AI truly serves humanity's best interests - is undoubtedly worth the journey.

Convening the Working Group on Investment in Responsible AI.

The journey from theoretical metrics to a practical, industry-wide benchmark begins with a crucial first step—forming the Working Group on Investment in Responsible AI (IRAI). 1Infinity Ventures will convene these working groups using the expertise of our GPs and their networks in early 2025. This diverse group of experts will bring together AI ethicists, veteran VCs, legal experts, data scientists, and AI startup founders. Their collective expertise will ensure that our approach is theoretically sound and practically applicable to the realities of the AI startup ecosystem.

The IRAI Working Group's initial task will be to validate the three metrics we've proposed for each of the five OECD AI principles. This process will involve rigorous debate and analysis, drawing on ethical considerations and real-world practicalities. For instance, when examining the Bias Mitigation Score under the human-centered values and fairness principle, the group might consider factors such as the diversity of training data, performance equality across demographic groups, and specific debiasing techniques in the AI pipeline.

Once the metrics are validated, the Working Group will define precise measurement methodologies for each. This step ensures consistency and comparability across different AI startups and technologies. The Environmental Sustainability Index, for example, might require the group to grapple with questions of scope: should it only consider the energy efficiency of AI models, or should it also factor in a startup's broader environmental practices? The Working Group would need to balance comprehensiveness and practicality in these definitions.

With clear metrics and measurement methodologies in place, the IRAI Working Group will then identify existing tools capable of measuring each metric, with a preference for open-source solutions. This approach would ensure transparency and lower the barriers to adoption for startups of all sizes. For metrics where suitable tools don't exist, the group might recommend developing new, open-source solutions to fill these gaps.

The next phase will involve testing these metrics and tools by measuring existing AI startups. This pilot phase would be crucial for identifying any unforeseen challenges or inconsistencies in our approach. The Working Group will select a diverse sample of AI startups at various stages of development and from different sectors, applying our metrics to these real-world cases. This process will reveal the need for stage-appropriate variations of our metrics or industry-specific weightings in our overall benchmark.

The IRAI Working Group must balance rigor and practicality throughout this process. While we need metrics that meaningfully evaluate AI startups’ ethical implications and responsible development practices, we must also ensure that our evaluation process doesn't stifle innovation or unfairly disadvantage smaller startups with limited resources.

The culmination of these efforts would be the compilation and publication of the first benchmark by the Working Group. This landmark document will provide a snapshot of responsible AI development in the startup ecosystem and establish a baseline for future comparisons. It would offer valuable insights into trends, best practices, and areas for improvement across the industry.

The publication of this benchmark will send a powerful signal to both the AI and investment communities, demonstrating a commitment to incorporating ethical considerations into AI investment decisions. It will catalyze a broader shift in how AI startups approach responsible development, encouraging them to integrate these considerations into their technologies and business models.

We are working towards launching the IRAI Working Group in the first quarter of 2025. With this, we are laying the groundwork for a fundamental transformation in developing, deploying, and investing in AI technologies. By bringing together diverse experts grappling with complex ethical and practical challenges and undertaking the essential work of real-world application, we're taking the first crucial steps towards a more responsible, sustainable, and ultimately more valuable AI ecosystem.

Establishing the Gold Standard

As we journey deeper into responsible AI investment, a crucial need emerges— a standardized, comprehensive benchmark that translates our previously defined metrics into a powerful tool for the investment community. We proposed an industry-wide standard, the Responsible AI Investment Rating (RAIR), maintained by an independent NGO—the Institute for Responsible AI Investment.

The RAIR would build upon the metrics outlined earlier, creating a holistic score reflecting a company's commitment to responsible AI development from an investor's perspective. This benchmark would encompass all the dimensions we've discussed: from the Social Impact Score and Environmental Sustainability Index under Inclusive Growth and Sustainability to the Bias Mitigation Score and User Empowerment Index under Human-centered Values and Fairness. It would incorporate our measures of Transparency and Explainability, including the Explainability Quotient and Algorithmic Transparency Score. The benchmark would also account for Robustness, Security, and Safety through metrics like the Adversarial Resistance Index and Data Privacy Protection Score. Finally, it would factor in Accountability measures such as the Ethical Governance Score and Incident Response Efficiency.

By aggregating these metrics with appropriate weightings, the RAIR would provide investors with a comprehensive score that offers a quick yet nuanced understanding of a company's ethical AI practices and their potential impact on long-term viability and risk. For instance, a company with a high Bias Mitigation and Algorithmic Transparency score might be viewed favorably regarding reduced regulatory and reputational risks. Similarly, solid environmental sustainability and social impact performances could indicate that a company is well-positioned for a future where consumers and enterprises increasingly prioritize socially responsible AI.

The independent NGO maintaining this benchmark would be responsible for refining the methodology for combining these metrics into the RAIR score, ensuring it balances ethical considerations with investment viability. They would collect and verify data from AI companies on each of these metrics, publish annual reports on trends in responsible AI investment based on changes in these metrics across the industry, and provide resources to help VCs interpret these metrics and incorporate them into their due diligence processes.

We create a comprehensive and actionable benchmark by basing the RAIR on these concrete, previously defined metrics. Investors can drill down into specific metrics that align with their investment thesis or areas of concern. For example, a VC focused on AI in healthcare might pay particular attention to the Explainability Quotient and Data Privacy Protection Score.

This benchmark could drive a "race to the top" in ethical AI investment. As Investors compete for higher RAIR scores in their portfolios, we could see accelerated funding for startups excelling in critical areas like bias mitigation, energy-efficient AI, and robust data privacy practices. The RAIR could become a valuable tool for limited partners (LPs) evaluating VC firms. An LP might look at the average RAIR score across a VC's portfolio to indicate the firm's commitment to responsible AI investment and its ability to mitigate ethical risks.

While establishing and maintaining such a benchmark presents challenges, including ensuring consistent data collection and keeping the methodology current in a rapidly evolving field, the potential benefits are immense. A standardized, metrics-based benchmark like the RAIR could bring unprecedented transparency and accountability to AI investment, helping to align the flow of capital with the principles of responsible AI development.

As we look to the future of AI investment, creating this independent, metrics-based benchmark represents more than just a new evaluation tool. It embodies a fundamental shift in how we assess the value and potential of AI companies, placing ethical considerations at the heart of investment decisions. By supporting the establishment of the RAIR, we in the venture capital community can fund the next wave of AI innovations and actively shape a more responsible, sustainable, and ultimately more valuable AI ecosystem.

A Call to Action: Join the Responsible AI Revolution

As we conclude this deep dive into responsible AI metrics, this is not just a 1Infinity Ventures initiative. We're calling on the entire venture capital ecosystem to join us. By collectively adopting and refining these metrics, we can create a standard driving the AI industry toward more responsible development.

To the founders reading this: We challenge you to incorporate these metrics into your development process from day one. Show us how you're building AI that's not just powerful but responsible. Demonstrate how you're considering the broader implications of your technology on society, the environment, and individual well-being.

To our fellow investors: Let's raise the bar for AI investments. Let's make these ethical considerations as fundamental to our due diligence as market analysis and technical evaluations. We invite you to join us in this mission. You can move beyond lip service to tangible, quantifiable commitments to responsible AI. By aligning our investments with these principles, we're not just mitigating risks but positioning ourselves as leaders in the next wave of technological innovation.

Another challenge lies in the potential conflict between short-term gains and long-term responsibility. In the fast-paced world of startups, there's often pressure to move quickly and worry about ethics later. Our role as venture capitalists is to demonstrate that responsible development and rapid innovation are not mutually exclusive. Considering these ethical dimensions from the outset leads to more robust, scalable, and ultimately successful AI companies.

To the broader AI community: We invite your input and collaboration. Help refine these metrics, develop standardized measurement methodologies, and create a shared vision of responsible AI.

And to all our readers: Stay engaged, stay informed, and keep asking the hard questions. Your awareness and demands for responsible AI will drive the market and push for continued innovation.

The future of AI is in our hands. Every line of code, investment decision, and product launch is a brushstroke on the canvas of tomorrow. Let’s ensure we’re painting a future we’ll be proud to inhabit—a future where AI enhances human potential bridges societal divides, and tackles our most pressing global challenges.

Together, we can build an AI ecosystem that is intelligent, wise, profitable, and profoundly beneficial for all of humanity.

The road ahead for AI is both exciting and challenging. As we witness advancements in AI capabilities, we must ensure that AI advancements are directed toward creating a more equitable and sustainable world. By focusing our investments and efforts on startups that embody the principles of responsible AI development, we can help steer the industry toward a future where AI truly serves humanity's best interests.

Whether you're a founder seeking inspiration, an executive navigating the AI landscape, or an investor looking for the next opportunity, Silicon Sands News is your compass in the ever-shifting sands of AI innovation.

Join us as we chart the course towards a future where AI is not just a tool but a partner in creating a better world for all.

Monthly Digest