The run up to the AI Summit San Francisco is well and truly under way and last week AI Business caught up with Robin Bordoli, the CEO of CrowdFlower Technologies. Robin has spent the past two decades helping high growth companies launch and scale platforms and products into rapidly transforming markets.
CrowdFlower is a pioneering data mining and crowdsourcing company based in Silicon Valley, San Francisco. Their ‘human in the loop’ platform provides businesses with high-quality AI solutions and a crowdsourced workforce. In this interview, Robin tells us how CrowdFlower is changing the game today and helping businesses find meaning from their data.
Human in the loop
The CrowdFlower platform trains, tests, and tunes machine learning models to make AI work. Their expertise provides a wide range of use cases that includes intelligent personal assistants, autonomous vehicles, medical image labeling, consumer product identification, content categorization, customer support ticket classification, social data insight, CRM data enrichment, product categorization, and search relevance.
In our interview, Robin describes the distinctive role of CrowdFlower further;
‘Our human-in-the-loop platform enables AI and machine learning teams to take unlabelled or unstructured data sets and enable humans (be they internal employees or accessed via some of our channels) to transform that data into high-quality training data. We do this by providing a number of tooling layers that ensure utmost quality, efficiency, and scale in the transformation of that data.’
As a forward thinking organization, we were intrigued to find out how Robin compares the impact of the internet at the beginning of the twenty-first century with the emergence of AI today; ‘It wasn’t just that the internet created a virtual crowd, but that the internet connected humans- breaking down physical and time barriers that have existed since time as we know it. The connections, combined with today’s level of compute power and data access has enabled the promise to take flight…
‘As transformational as the advent of the internet was, the potential of AI is exponentially greater. It’s not just about connecting (a finite (albeit vast) group of humans, but about accelerating and even automating the infinite world of discovery.’
Robin continues; ‘granted, there’s a tremendous amount of heavy lifting that has to happen up front, when you stop to think about what our world will be like with the acceleration of discovery, it’s absolutely mind boggling.’
Robin outlined some of his predictions for the regional impact of AI adoption over the next 5 years and the challenges we need to anticipate.
‘In Silicon Valley we see a lot of what is going on right in our backyard and these local efforts alone are transformational. But given our connected world, innovation is happening in every corner of the planet. It’s not just the US and China, but Europe, all of Asia and even other places like the Middle East. I think rapid growth will happen globally.’
‘Where there is anything of value, protection always becomes an issue. This used to be more focused on physical goods, but as the world becomes more digital, data protection has started to eclipse the issue of protection of physical property. This will continue to be the case.’
‘With the advent of AI, the issue is not just the data itself, but the way in which the data is structured to teach and train models. If you think of ML models as empty brains with infinite potential, then they need to be ‘taught’ (with human labeled data) to disambiguate between things like ‘appropriate or inappropriate content’, ‘fake news from real news’, ‘a suspicious character from a regular citizen’. Human subjectivity- and bias introduced into training data then become hard wired into an ML model autonomously moving forward. These ethical issues are some of the biggest issues facing the AI community today.’
‘AI for Everyone’
We closed the interview by discussing the vision behind CrowdFlower’s AI for Everyone challenge;
AI for Everyone is based on our belief that the power of AI should be accessible to everyone and that financial backing should not be a barrier. The challenge takes the form in the value of a million dollars earmarked to allow 8 projects over the course of the next year to move accelerate their projects with our services. We actually have selected our first two winners and will be announcing them soon. Stay tuned.’