Tabitha Goldstaub of the UK AI Council and CognitionX elaborates on why AI must be human-centric for its full potential to be unlocked.
At the end of April of this year, the United Kingdom announced the creation of a new Government Office for Artificial Intelligence to help ensure the nation is positioned on the leading edge of AI technologies. As part of that announcement, a new AI Council will be created to advise the Government and the Office of AI to promote the adoption of the AI across industry, government and academia.
Tabitha Goldstaub, co-founder of CognitionX (an AI advice platform), was tapped to serve as the chair of the UK AI Council. Tabitha was also selected to lead as the AI Business Champion as part of the larger Grand Challenges technology program, announced by Prime Minister Theresa May. Recently, we sat down with Tabitha to discuss some of the prevailing myths and realities concerning AI and her plans to help promote business investment in AI research and development.
“Data is key to an AI product successfully outperforming another, but the potential competitive edge that excites me most is how well a product supports a human-in-the-loop model. Computer-human interfaces and the right tools for collaboration will be a big win. AI systems that support people to do their jobs better and more enjoyably should be the systems that rise above the others.” – Tabitha Goldstaub, UK AI Council and CognitionX
ANSWERS: You’ve recently been announced as the AI business champion and chair of the new AI Council for the UK government’s Office for Artificial Intelligence. Can you describe for us what your role entails and some of the AI-related challenges you’ll be addressing?
TABITHA GOLDSTAUB: The AI Council is here to help coordinate and grow AI in the UK. We’re an open forum for collaboration and problem solving between industry, the public sector and academia. We’ll be advising the government and the Office for AI, and we’ll also be looking to improve the understanding of AI across the UK, identifying any barriers to growth and seeing where we can support.
There will be three immediate focuses that are already set out in the UK government sector deal on AI. The first (skills) hinges on understanding what’s needed and improving the number of people who are able to be a part of this new wave of technology. At the same time, we want to promote a sensitive approach to data and data ownership. The third, around adoption, will concentrate on making sure that both small and large companies can start using AI safely and confidently. Along with these three areas, we will also look ahead to the longer term, more strategic needs of the UK.
ANSWERS: Obviously there’s much interest out there in AI, as evidenced of course by the formation of this Council and programs like it. What is the biggest misconception that you regularly encounter about what AI can do for an organization, and conversely, what are the realities?
GOLDSTAUB: The biggest misconception is lack of understanding of the scale, breadth and pace of change that we are undergoing. Many people still have little idea of how big this paradigm shift really is. The reality is highlighted by the investment being made by the tech giants and countries themselves; China for example is investing $425 billion in AI by 2020.
ANSWERS: What do you see is the key to encouraging businesses to invest more in AI research and development?
GOLDSTAUB: I think the most important thing is to paint the vision of where they can get to. There’s a quote attributed to Antoine de Saint Exupéry (the author of The Little Prince) that I think sums it up:
“If you want to build a ship, don’t drum up the men and women to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea.”
And I think that that’s what we need to do with businesses; they need to see the value of where AI can take them rather than the nuts and the bolts of the details. If businesses can see that value, then hopefully we can start to also dispel the fears and the concerns.
At CognitionX we do that by showcasing great stories. There are really good case studies out there where AI has been used in specific areas. I know every time I hear them I say, “I wish I could do that.” That’s the emotion we need. For example, Unilever reduced the time to hire of their graduates from four months to three weeks using AI, creating a much better candidate experience. What really excites me is that by automating the drudgery work, it freed up talent advisers to spend more time with candidates, which led to much happier staff and new hires.
We really need to empower people who see these great stories to feel that they can achieve this, too. CognitionX offers an expert network where we connect the people who have questions with the people who have the answers, and we try and make it as collegial as possible, so that they don’t feel like they’re alone. They’ve got a global network of people at their fingertips who are also trying to deploy AI responsibly.
ANSWERS: How do we ensure and promote transparency in AI research among development groups when you have players out there that may be looking for a competitive, first-to-market edge?
GOLDSTAUB: There are many elements to transparency, from technical transparent AI systems to transparent business practices.
I’m encouraged by the creation of the Centre for Data Ethics and Innovation, which is being led by Roger Taylor. The name says it all: ethics and innovation have to come together, and hopefully this group will aid an increasingly transparent approach to AI research. The Centre will advise government on the measures needed to support and enable safe, ethical and innovative use of data and data-enabled technologies.
At CognitionX we track the developments in responsible AI so that businesses can get a better understanding of what’s being built. For example, there’s an interesting project from DARPA (Defense Advanced Research Projects Agency) called XAI, which aims to produce a glass box model rather than a black box model. XAI aims to provide an explainable way for the humans in the loop to understand what’s going on without sacrificing AI performance. It’s early days, but it’s an interesting thing to see.
ANSWERS: Do you think we will see increasing data sharing partnerships between organizations to create more robust training sets for development? If yes, how do you think consumer privacy concerns such as the GDPR will be managed in that kind of environment?
GOLDSTAUB: Access to more data is one of the key enablers of this AI revolution. It definitely wouldn’t be possible without the creation, annotation and labeling of new datasets, whether they are university datasets or from organizations themselves.
And so, we’ll see increased data sharing in many different ways. One is through commercial partnerships. It’s clear that proprietary datasets will give companies a commercial edge. But we are now seeing companies come together to actually work out how can they collaborate, use similar datasets and move things forward. That’s why I’m excited about the UK government’s data trust proposal in the sector deal which looks at exactly this.
The second type of data sharing we’re seeing is through open data and government-led initiatives – for example, with Ordnance Survey and Geospatial data.
The third is where consumers are opening up their own data. We’ve seen data transfer partnerships (Google, Twitter and Facebook) come together and make a way that consumers can actually port their own data across these systems.
I think the good news is that we’re seeing an increased focus on how we safely make more data available. We don’t have all the answers, but the direction shows that all parties are dedicated to having this discussion.
ANSWERS: As more and more organizations embrace artificial intelligence, what do you see will become the competitive edge if such technologies become ubiquitous?
GOLDSTAUB: Data is key to an AI product successfully outperforming another, but the potential competitive edge that excites me most is how well a product supports a human-in-the-loop model. Computer-human interfaces and the right tools for collaboration will be a big win. AI systems that support people to do their jobs better and more enjoyably should be the systems that rise above the others.
ANSWERS: What excites you the most about what AI means for business in the coming years?
GOLDSTAUB: This is a critical moment in time, and those who quickly understand where AI can be used across their business will benefit most, if they consider how to do this responsibly, ethically and safely. It’s an exciting time to re-look at what we’ve already created in the tech world and say, “Is this what we want? Is this how we want to reap the benefits of this incredible technology? How can we make this more accessible? More beneficial for more people?” So now is the time to reevaluate and build technology with AI that works for everyone in society, and I’m excited about that.
In our new series, AI Experts, we interview thought leaders from a variety of disciplines — including technology executives, academics, robotics experts and policymakers — on what we might expect as the days race forward towards our AI tomorrow.