As part of the ongoing interview series kick-started on Women’s Day 2021, Team IA brings you insights of women bosses who inspire with their grit, determination, formidable achievements, and a brilliant career trajectory, owning one goal at a time.
Team IA gets chatting with Jennifer Jordan, Managing Director at Techstars for Barclays’ New York Accelerator and Female Founders First program.
As a VC and Vice President of MassVentures, she led the firm’s investments in Ginkgo Bioworks (synthetic biology), ClearGov (data transparency), Spiro.ai (ML and AI that lets salespeople sell), MachineMetrics (industrial IoT for advanced manufacturing), and FairMarkIT (digital procurement).
She is currently focused on an investment thesis related to the tools and data required to bring Trust & Transparency to AI. She is dedicated to ensuring: that female and diverse founders have access to the networks, advice, and capital they need to thrive; that women and diverse investors gain decision making authority in VC to close the Gap Table; and, that enterprises adopting AI to automate decisions benefit from its full potential because they have the data and tools available to make sure everyone is represented.
What are your thoughts on the percentage of working women professionals in the field of AI, especially since you are an early-stage investor and get to meet those driving businesses more often?
In the US we have a couple of interesting trends. Firstly, after a long decline from a peak in 1983, when women were earning 37.1% of bachelor degrees to a low of 17.6% in 2006, the number is increasing again reaching 21.7% in 2017. I want to see that number increase even more. But what is really interesting is that women earning advanced degrees, masters, and doctorates in engineering, CS, and Information Science continued steadily up and to the right in that time so that in 2018 they earned 40% of the Masters degrees and 50% of PhDs. In the field of AI in 2018 women earned 25.5% of the Masters and 19.4% of the PhDs. What makes this so striking it is that it means we could be doing so much better in the private sector. Looking at research published by researchers at major AI first companies, Google, Microsoft and IBM, respectively only 11.3%, 11.95%, and 15.66% were women.
Where are the women in AI?
I see women in AI doing three things: 1. They are founding the companies, they want to see in the world such as Rana el-Kaliouby founder and CEO of Affectiva, or Beth Porter of Riff Analytics, Navrina Singh of Credo AI, Marina Pavlovic Rivas of Eli Health and Alicia Chong of Bloomer Tech or; 2. They are advocating for the accountability they want to see in their field and working to represent and protect those least represented in the data such as Joy Buolamwini of the Algorithmic Justice league and her co-author and former Google AI Ethics researcher Timnet Gebru or; 3. They are driving advances in the field and raising the next generation with their academic work like Rosalind Picard, Regina Brazilay, or Julie Shah of MIT.
Since you yourself are a change driver who impacts and imparts large-scale decisions, what sort of changes are being made in AI for the greater good of women?
AI has the power to transform and the power to inadvertently harm if we do not pay close attention to the data going into the system and the predictions coming out. Those who are least represented or represented but least able to advocate for themselves will be the most impacted. At the same time there is enormous opportunity in the developing world to use AI for Good. For example, the Botnar Foundation used an AI model in the field to help triage at risk pregnancies use telemedicine, and make sure the women with the highest risk were the ones who reached hospital care. It offers tremendous scale. But we need to do better. Alicia Chong’s company Bloomer Tech was formed to build a women-specific medical grade heart monitor because 1/3 worldwide are at risk for heart disease. 54gene, while not female-backed is building a company to make sure that as we move forward with genetic research to solve problems like understanding how to fight Covid-19 the data sets we use to develop diagnostics, vaccines and treatments do not rely 87% on the genes of only white men of European decent.
You drive in millions of dollars across companies. How does it feel to be a leading professional in the field of AI?
I am an early stage investor – a generalist with a focus on data and tools for AI Trust & Transparency – so it is my privilege every day to meet founders who are driving to make their vision a reality and to solve the problems they see in the world. It is my joy when I get to back amazing founders – especially female and diverse founders tackling venture scale problems that they are able to see and solve especially because of their unique experience of the world.
You are known to spend quite some time on AI Trust & Transparency, a cause very dear to many AI professionals. What are some of the hurdles faced in building this trust, and how could we overcome them?
Trust is a fundamental human issue – with more than one side to it. We trust ourselves and others to make decisions, even though we are prone to commitment bias and inherent bias. And at the same time, we have relatively little understanding of the amazing complexity of our brain that lets us process 100,000s of thousand of signals to listen to a bird sign and follow the sound or feel the hair stand on the back of our neck when there is a threat, or still find our way through the house with our eyes shut. People who are AI ‘purists’ say ‘how can you say you need to have trust or transparency in AI when humans can’t explain all their decisions.’ We can’t explain all our decisions, but we spend a lot of time trying. And with AI I think that it is the point. What tools, data, and processes can we implement to improve our visibility and understanding of what an AI system or a model is intended to do, what data did we input to train the model, where did it come from, are there biases inherent in the data set or gaps, are the outcomes what we intended, will the model still make accurate decisions over time or have the inputs or the population it was intended to decide about changed (drifted), is it secure, does it learn. These are hugely important things to have some process and visibility around because we are talking about situations where decisions can be made at an incredible speed on an enormous scale. Facebook, Google, Amazon are using AI to make billions and billions of decisions a day about which media to show us, what search result to display, which product to show and which price to offer. Sometimes the systems scale unintended optimizations like Youtube’s algorithm accidentally showing young girls pictures of friends on the beach to 100,000s of paedophiles. Another example is that of Apple and Goldman Sachs releasing a new credit card that was biased in how it determined the credit limits for women. No company wants to be in that place and a lot of AI investment right now is held up from going into production and delivering its return because we have not been building trust and transparency from the outset. In fact, this is a problem that only continues to grow as the use of AI grows, wherein we have models feeding decisions to other models. So the market for tools for AI trust and transparency will grow exponentially as use of AI grows – I think it ends up looking like the cyber security market – growing from about $5B today to some $50 billion in 5 years.
What does it take to bridge corporate and entrepreneurial disciplines? Can AI help here too?
This question makes me think that it might be awesome to have the ability to model an organization, look at its competitive landscape, talent, core capabilities, financials, and run that across the data we have on emerging companies to see which innovations, investments would provide us the greatest advantage and opportunity. But more modestly, there are tremendous opportunities to leverage AI throughout organizations from automating processes to enable people to do higher level work, to building new models that better understand what products to build or what services can be created to meet customers’ needs and drive new revenue growth.
It’s a terrific time to be a founder or investor working in AI, but it we want to see it deliver on its potential we need to build for trust, transparency and accountability and we need to be sure we are engaging with a diverse, inclusive, global set of employees founders and researchers because a rich differentiated experience is needed to make sure we are thinking deeply about what we build, and its consequences as well as rewards.
Thank you Jennifer for the insightful interview. We at IA are thrilled to bring to you more engaging interviews. Watch this space and do subscribe to our newsletter.
US Department of Education report on Degrees in computer and information sciences conferred by post secondary institutions 1970-2018.