Investing in Ethical AI

Ethical Intelligence: How did you come to this focus on responsible tech? Is it something you set out to do from the very beginning, or one that has grown over time?

Sarah Drinkwater: I began working in the technology industry in the late 00s when there wasn’t a developed notion of the power the field held. I’d grown up with a self-taught developer dad and been a very early internet user, so in my household, there was a lot of idealism around the possibilities tech brought which echoed so much of what I felt as a startup and then Big Tech employee.

But after time spent in product teams, it became clear to me that, despite our good intentions (I rarely saw otherwise), mistakes were still made. Intentions weren’t enough. I then moved to run a space for entrepreneurs in London at the same time the techlash began bubbling up; there was an oppositional dynamic brewing between the tech community and the media, who had previously adored the industry and now began asking questions. I left Google to join Omidyar in early 2018 to establish their responsible tech team.

EI: How have you seen the responsible tech space grow and mature over the past five years?

SD: When I made that move, it was seen as a pretty wacky decision but I had absolute faith in both the responsible tech scene becoming a thing but also the idea of responsibility becoming a powerful moat in driving customer trust, employee love, and more. It’s been very satisfying to see this belief come true; we now have multiple household name companies with ethics boards, job boards like All Tech is Human, specific conferences such as Women in AI Ethics and a plethora of interesting jobs that help to operationalize all the work that lives below principles.

And outside of those who work in the field, we’ve seen their tech worker colleagues hold them accountable through the tech labour movement, we’ve seen LPs and VCs increasingly ask deeper questions about practices (but not enough yet) and we’ve seen companies who refuse to engage in ethics washing and take hard choices win. 

EI: You’ve worked with entrepreneurs, investors, and even tech workers on responsible tech, so it’s safe to say you have an in-depth view of the different perspectives in the ecosystem. How are tech workers driving the motivation and need for ethics in technology?

SD: When I first began at Omidyar, we spent deep time thinking through the tech ecosystem and where intervention points were for us to fund. As long term impact investors, we obviously started with VCs and LPs but, as a former tech worker, I kept thinking about how hard to hire certain roles are. You look at the narratives of many of the largest firms employed in the 00s, focused on their incredible utility, and how proud workers felt to work for X firm. That’s changed in the case of certain big companies, at the same time as those firms have grown, requiring more talent, and at the same time as there are many more places those workers can go. Shame is an incredible incentive; at its very core, people want to do the right thing and be proud of where they work. So, whether it’s advocating internally or voting with your feet and leaving, tech workers are essential in building the right systems. 

EI: What role do entrepreneurs play in developing responsible tech? 

SD: Systems might be built and held accountable collectively, but so much of the culture and direction of a company comes from the top. Salesforce made the first big tech company move on creating the chief humane use officer because Marc Benioff cared. But, beyond that, there are multiple choices made early in company creation, from business model to equity pool to choice of investors and partners, that can make a real and true difference. 

EI: What can investors do to support entrepreneurs and tech workers in the growth of the responsible tech movement? 

SD: My core concern is that the venture model only works when companies are exponentially successful, and it’s hard for companies to scale so large without systems breaking or shortcuts taken. The right investor can be an incredible steward and help the founders and teams avoid this, but it’s on them also to recognise that sustainable growth can be as exciting as exponential growth. 

EI: Now as an angel investor yourself, how would you describe your approach to investing in technology?

SD: I’m early in my journey but I was excited by the chance to work first-hand with founders, as I did at Google and then at Omidyar, in helping operationalize the work of building responsible companies. I’ve tended to focus on companies with community-directed products, which is what I know best, with diverse founding teams and rock solid values alignment. 

EI: What role do values play in this approach?

SD: Values are critical, and I suspect not enough founders or investors spend enough time getting these down on paper. For me, there are simple things such as: I don’t invest in ad-based models, or businesses that are free to users but rest on selling data. I like products that give both utility and delight. And I’m particularly interested in companies with strong governance, especially if that’s shared with their

community - such as Anyone with their external ethics board which you convene or Library of Things with their innovative governance structure. 

EI: How does a startup’s mission and values impact its success?

SD: In Mind Foundry’s case, where I’m an non-executive director, they’re working on an important mission; real world AI for high stakes environments such as the public sector.

“It would be impossible to do that work well - to build the necessary trust with partners, or to hire the very specific in demand talent needed, without a very strong mission and values.”

The AI Promise

AI is consistently cited as one of the top macro-trends poised to dominate the next decade of innovation. Chatbots, conversational systems, banking apps, smartphones, self-driving cars and even home devices, like Alexa, have rapidly become part of our daily lives. As intelligent automation technology becomes increasingly pervasive across all industries, McKinsey Research predicts it could unlock $3.5T - $5.8T in annual value, roughly equivalent to 40% of all data analytics techniques.  

This promise has unleashed an unprecedented level of VC investment in AI.  An OECD study analyzed VC rounds in 8,300 AI startups worldwide, covering transactions between 2012 and 2020 that were documented by capital market analysis firm Preqin.  According to the findings, the global annual value of VC investments in AI startups grew from $3B in 2012 to nearly $75B in 2020 with startups in the US and China capturing over 80% of the investments.  The EU followed with 4%, trailed by the U.K. and Israel at 3%.  In October of 2021 it was reported that 78 AI startups had reached unicorn status (exceeding valuations of $1B) including Argo AI, Anduril, Open AI, 6Sense, Preferred Networks and many others. 

AI at a Crossroads

In spite of the promise of AI, many view the technology as a double-edged sword.  Consumer attitudes towards the development of high-level machine intelligence is marked by fear, doubt and skepticism. The public is evenly split on whether this is a positive or negative development for humanity, or whether these technologies should even be developed in the first place. Confidence falls much farther when it comes to trusting major tech corporations like Facebook, Amazon, Microsoft, Google, or Apple for the development and stewardship of these innovations. While Skynet remains the stuff of SciFi novelty, the contemporary mainstreaming of AI across a broad variety of industries, from healthcare and financial services to autonomous vehicles and manufacturing, is now driving a unique set of legitimate concerns around Ethical AI Governance. 

Our Perspective

As an industry, we are only just beginning to understand and develop best practices for AI Ethics and governance. This knowledge is, and will continue to be, essential in accelerating the adoption of responsible and trustworthy AI. At EAIGG we believe this process of developing best practices and knowledge bases should be a collaborative initiative, bridging  startups, investors and corporate enterprises in a joint effort to make these resources openly accessible for the betterment of the AI industry as a whole, and society at large.  Against this backdrop, it is incumbent upon AI practitioners and executives to lay out a shared set of principles, frameworks and standards to apply to AI technology development, deployment and operations to mitigate and address ethical AI governance concerns. The Ethics Maturity Continuum is the first of such, an openly accessible foundational framework designed to meet this growing need. 

Because venture capitalists occupy a unique role in shaping the arc of innovation, we unveil this framework primarily as a due diligence tool that investors can use to evaluate a company’s health with regard to the ethical development and deployment of AI technology solutions.  We believe Ethical AI will soon receive similar treatment as ESG issues in the public eye.  Venture funds will have to answer to their LPs, and to various stakeholders, when justifying their investments, and this tool provides a practical diagnostic health check to evaluate those decisions.  Business executives may also be interested in evaluating their own progress when it comes to these issues, to assess their progress over time, and to diagnose areas of improvement or of excellence.  As AI development touches a broader array of stakeholders, we anticipate an increasing need for ongoing due diligence in these critical areas.

Maturity Continuum Research Background

"AI ethics is a set of values, principles, and techniques that employ widely accepted standards of right and wrong to guide moral conduct in the development and use of AI technologies." 

- Alan Turing Institute

In recent years, the impressive advances in the capabilities and applications of AI systems have brought the debate  of AI’s impact on society into sharper focus. In order to maximize the benefits and mitigate the risks of these new solutions, the field of AI Ethics has quickly emerged, propagating new tools, frameworks, and Ethics as a Service products designed specifically to ensure that algorithms are built with ethical principles at the core. 

To aid in this adoption, the majority of global technology companies have established ethical frameworks and policies, while multiple governments have released similar high-level frameworks. As enterprise and regulation begin to mandate  ethical technology practices, an imperative has surfaced for startups to understand how to build these practices into the company and technology from the very outset. 

The AI Ethics Maturity Continuum consolidates and synthesizes these global frameworks and policies to determine the key factors for ethical technology development for early and late stage startups. These components have then been translated into a comprehensive model that provides a method for assessing a startup’s ethical development maturity along with clear actions points for improvement.

Benefits for Startups

Responsible AI, when used to its full potential, is a source of competitive advantage. Building ethics into technology and product management has a number of business benefits, including:

• Enhanced Product Quality - Incorporating responsible, independently verifiable AI practices throughout the product development lifecycle results in better alignment with customer wants and needs.

• Employee Retention - Ethical practices significantly benefit talent acquisition, retention, and engagement, especially in the current competitive labor market.

• Sustainability - Focus on value-sensitive design leads to better data management, which reduces digital waste and increases sustainable practices.

• Regulation Readiness - Companies must stay ahead of AI regulation to remain competitive and active on a global scale, particularly where innovation eclipses governmental action.

• Increased Growth -Ethics improves top-and bottom-line growth by increasing customer engagement, broadening revenue streams, and offering procurement advantages in competitive bidding processes.

• Brand Loyalty -Heightened focus on the sector has increased the importance of customer loyalty to brands consumers believe in, and trust.

Benefits for the Wider Community

In addition to startups, this framework is also a valuable tool for larger enterprises and customers. For large corporations, it enables assessment of ethical practices within a prospective company during an acquisition process. For customers, it creates a transparent process to analyze the values of the company providing them with AI solutions, which in turn creates customer trust in the product. Overall, the application of this framework stimulates confidence in the adoption of responsible AI technologies for the community at large, fulfilling the promise of these breakthrough technologies while mitigating the risks.

When we first started,  my CEO Chris and I figured this seed funding journey was going to be a rather smooth one. We were both confident in our mission. We were solving the challenges of almost every large organization – wanting to do the right thing in the data driven AI era and reporting to stakeholders and customers about it. We figured that the ESG’s, Taxonomy, and upcoming EU-regulations on top of the exploding social concerns were leading organizations to aim for transparency and reporting metrics on their ethical AI. We were paving the way for organizations pioneering transparency, and tools evaluating the ethics of AI.

That might be the case, but our journey was difficult. We started to blame VC’s for not “walking the talk” on responsible investments. Investors still seemed to prioritize trade-offs between profitability and sustainability, despite increasing concerns over trust and the push towards responsible AI. They appear to lack the understanding of the ethical and societal pitfalls and risk exposures in AI driven tech scale ups.

Without using an ethical filter, they are investing in the companies lacking long term and sustainable business models. 

Because of investors' lack of understanding of the ethical AI space, we failed to explain to the receivers of our pitch deck and presentations what a sustainable AI business model means, and how we create value meeting the next generation’s commitment to responsible and trustworthy technology. We failed in our initial attempts to explain that we are at a crossroad in the AI industry. 

Eventually though we did succeed in finding our lead investor, BGV, that truly aligns with our core values and business strategies, and most importantly understood the opportunity of responsible tech. 

We learned a lot over our seed funding journey, important insight into investing in Ethical AI that I now want to share with you. 

AI for sustainability is not the same as Sustainable AI

The lack of understanding in the investment space meant that we were unable to explain to VC’s the difference between AI for Sustainability and Sustainable AI.  Typically, VCs engaged in responsible investments screen tech start-ups for AI solutions targeted at achieving one of the UN SDG’s. But there are responsible investments which don’t clearly map onto these. At anch.AI we are meeting these SDG’s indirectly, but it is not our business model. 

What is the difference between the two? Well, you can have an AI solution for sustainability, without it being sustainable AI. For example, an AI solution for crisis management in distressed conflict areas might have an algorithm specifically helping mothers and children get information and access to shelter and aid. However, the algorithm is built on data that is risky to obtain and secure, as well as takes mass amounts of energy to maintain. This would be an AI solution for sustainability but without Sustainable AI. 

Most VCs are still in the tech silo-perspective when it comes to ethical AI governance

We are introducing a cross-functional approach to the ethical AI governance landscape by being an orchestrator of ethical AI. This means activating the tech, legal and business teams to align on critical ethical and legal considerations and trade-offs. This is our unique differentiation. 

Why is this important?

For example, fairness and explainability have different meanings based on if you are working in the tech, legal or business teams. Lawyers are not trained to understand code; coders are not educated in the organizational ethical values and principles, nor can they code them in coherence.

Operating in a technology silo means opening the door to costly societal risks and legal breaches.

We failed to explain that existing AI governance platforms claiming ethical AI governance only assess and govern AI from the tech silo/perspective, which gives a false sense of security. Because investors did not understand this gap, they were unaware of the need for such a tool.

Lagging VC screening tools

You can have fast scaling and huge profits in a short time. Black box AI solutions or AI solutions with mitigation tools from the tech silo/perspective will fail to dismantle ethical risk exposure in a short time. With this perspective VC’s will be exposed to fail financially in the long-term perspective. Our platform would be a strong tool for VCs to screen their investments for exposure to ethical risks and ensure their own business model to be sustainable. 

In the middle of our seed investment journey, I got a call from Anik Bose, general partner at BGV. He did not know we were in the middle of raising seed funding, he contacted me to invite me to the non-profit Ethical AI Governance Group about to be founded in Silicon Valley. We ended up discussing a set-up for BGV to be lead investors.

For the BGV team we didn’t fail to explain anything, they already knew from the start, educated by Olivia Gambelin, that our product was strong and that the market for ethical AI governance was about to grow fast. BGV truly understood the responsible tech space, they could spot a trustworthy and ethical product, and understood the value of our product. The BGV team are truly committed to ethical AI governance combined with knowledge and experience in B2B SaaS. 

Just how important is the first investment?

One of the biggest mistakes a founder can make is thinking that securing the initial investment is an immediate win. But it’s not just about the money; investors aren’t merely cash flow, they’re partners who are going to stick around. And just as with any partnership in life, you want to ensure this one is built on foundations that will last. However, it can happen that either side will want to take the company in a different direction, resulting in founders finding themselves in the difficult position of compromising on their vision and investors worried about the strength of the founding team.

The initial investment will have a huge impact on the future of the company, so it’s essential to take the time to find the right investor.

But how? And what can we do to avoid investor incompatibility?

Ethics: The Unexpected Solution

A significant part of investor incompatibility is simply that investors and founders want different things. One might for instance favour sustainability over profit, or scalability over user experience; even simple decisions over who to hire and fire can contribute to building tension. These differences in values and priorities create frustrations between investors and founders. So finding investors who prioritise the same values, expectations, and visions as yours will help to avoid investor incompatibility.

It’s that simple - understanding your own values, and making sure that key values are synonymous within your investors means you’re just less likely to disagree on key decisions.

Why ethics benefits founders

Understanding your ethics and values as a founder not only reduces investor incompatibilities, but it can create investment opportunities from the outset. 

How? Understanding your ethics and values shows in conversations with investors. It can give you a clear idea of how you want to see your company’s mission come to life and the ability to speak confidently to it, which in turn attracts investors that share a similar vision.

What does value alignment mean for investors?

Investing in companies that align with your values means you are investing in the long-term success of a company, founding team and project you believe in. From the outset, you will have a strong understanding of the issues that matter to the company and how a founding team will approach solving those issues, giving you further confidence in your investment.