 So we're closing out SOCAP 23. What an absolute privilege and I'm sure you've all had Wonderful week same as I have so as introduced. I am an impact investor I work with Caterpult and we invest in early-stage impact technology companies focused on climate ocean and food and agriculture in Africa And we also help to raise awareness through the Caterpult Foundation about the tools of impact investing and impact technology for good And more recently I am a master's graduate from the University of Cambridge After a big 20-year gap I went back to embrace academia and to put myself through the Troubles and Turbulations of going back to school to really try and understand this world of AI ethics and How I could bring that into my work at Caterpult with the founders and with investors that we work with See what overlap there was and how we could learn from AI ethics in our work as impact investors So I'm going to share some of that research now in eight minutes and 55 seconds bear with me But first off I'd like to buy a show of hands who has Companies in their portfolio that are using artificial intelligence machine learning as a value driver Okay, super quite a few as Expected and this is sort of what I found in my research So I looked at pitch book database and analyzed portfolios of 350 impact investors and found over 2000 deals in their companies been invested in by impact investors across all of the SDGs that we're using AI on machine learning Which is not surprising given that the numbers continue to arise that have been invested into AI and And Goldman Sachs are telling us it's not a bubble So we anticipate that this will continue and research is also showing us that it can be used for solutions to our environmental and social problems and the stat I quote here is from academic research showing it can support 79% of the UN SDGs But what does that mean for us as impact investors? There's lots to get lost in there's a lot of hype There's a lot of noise AI for good AI for bad. How can we navigate our way through that? So this hence my interest in AI ethics. So by another show of hands How many of you that have looked at an AI have thought about integrating AI ethics into your work? Oh Beautiful Okay, super. So you're proving out my research Thank you, so cap in that there are more that have done the investments and have started to think about sort of AI ethics But I'm quite impressed at this little group here. You need to come find me afterwards But what I did find from interviewing 14 impact investors both private wealth and impact VCs That there is definitely sort of an interest in learning more and Recognizing that an integration of these themes of AI ethics can be supportive in Achieving sort of their goals of being making better impact investments So what does that mean sort of for us and sort of where do I sort of go next with this research? And what are these risks that I'm talking about? Well, these risks are real First off just thinking about techno-solutionism, you know critiquing ourselves is technology always the best Tool for the solution to the problem that we're seeking But then thinking about some of the investments we might see so examples such as a Company seeking to debias the recruitment process by using sort of algorithmic decision-making But if that tool has been built on data that was already biased We're just going to perpetuate sort of existing biases and we're not going to get sort of a de bias recruitment process or a health company a mental health company using classifiers looking at language being used by individuals to see and raise awareness if somebody's at risk of self-harm But again, how has this tool been built? How do we know it's classifying and it is going to sort of escalate the correct Issues those that really are at harm and who might get missed So by asking these questions about how the tools have been built and thinking about sort of the ethical considerations of those tools Helps us think about sort of the overriding sort of impact of those investments So the second part of my research was to think about Okay, what do we know what is out there within the AI ethics world? What work has been done that we can integrate into our impact investment? processors and In talking to the impact investors This is what I looked at Our investment processors. How can we think about different steps and stages in our processors? That could lend themselves to integrating AI ethical considerations This was my hypothesis and I tested this with those that I interviewed So first off screening due diligence Some of what I just flagged as high level. How can we get those questions? How can we as impact investors integrate sort of more effectively the right questions to ask? So we can get underneath sort of how the companies are using these models building and deploying these models to ask questions That are going to illuminate for us what the potential AI risks may be And then once we've done that, you know, there are tools out there So there are tools out there should mention some of them are here and there are many others We have Ravid Dottan who I work with who has put together a framework to support investors on diligence We also have NIST here in the US with their risk management framework There are things out there that can help us and can be used and adapted for our processors So once we've asked those questions and we have our answers Contracting I'm sure many of you have seen I have in my time working in impact investment this Increased number of us that are integrating impact terms into our term sheets into our contracts How could some of those be adapted? how could we sort of add in and Cover some AI ethics Terms within those contracts. So for example, you know Embedding sort of the purpose and an AI tool can be used for multiple purposes But if we're investing for the energy efficiency purpose, or we're investing for sort of the health outcome How can we embed that and ensure that the company and hopefully ensure that to support the company to head on that Trajectory and not to use a tool for something that goes against the impact that we're looking for So contracting being the second process it and then the final one Follow up show up use our voices. We sit on the boards of these companies. We mentor them We run accelerator programs for them Ask them to report Give us the feedback loops so that we can get the data as to how these models are working and what are the risk management processes that the companies have put in place ask for that data support the companies and Be sort of a good steward for those that we invest in from an AI ethics standpoint So this is what I proposed and this is what I spoke to my Interviewees about and I got some resounding yeses. This sounds great. Alison. Let's sort of go out. Let's go build it So that's sort of the next piece and I want to sort of you know I'm not an academic but I'm an action-orientated impact person and so you know in distilling sort of my two years and 16,000 words Dissertation the key point I'd like to make here today is that this investment is continuing There is something that we can do and if it's not us then who is going to sort of take that step So you've heard from me and I've done quite well. I've gone pretty fast and I've got one minute and 51 seconds left So now I'd like to hear from you so this is sort of my call to action from the end of this sort of talk and That I'd like to learn from others that have started those of you here that raised your hand that have started thinking Reach out to me. Follow me on LinkedIn. Follow catapult foundation Revita and I are actually doing a webinar next week on her work and we'll continue to do so We will sort of put together sort of content on this topic over a period of months and we want to sort of bring those of us together that would like to learn more so that we can put together some best practices and I would also like to build out further the tools that we've considered and make them publicly available So the catapult foundation supporting this but I'm also looking for other funders so that it's not just me I can sort of work with others and we can build out sort of more of these tools So I hope and by doing this that next time if I come and ask those to put their hands up that Are working with AI ethics integrating AI ethics. We have a few more of you in the audience raising your hands Thank you very much