Key Takeaways We Learned at The AI World Forum

AI World Forum Summary
By: Digital Finance Institute | 03/11/2019 |Last Updated: 03/11/2019

(From Right to Left) Peter Suma, Manir Al-Faisal, Jean-Baptiste Passot, Anjalee Narenthiren, and Jason Braverman on the Robotics Panel at the AI World Forum hosted at the Carlu in Toronto on March 4-5, 2019.

Last week, the second annual Artificial Intelligence World Forum was held in Toronto where global thought leaders discussed AI, Machine Learning, Autonomous Vehicle, Infrastructure, Banking, Media & Entertainment, Healthcare, Robotics and the major issues we are facing with AI such as Diversity in Coding, Ethics of AI, and Cybersecurity. For two days, great minds from Uber, Shopify, MasterCard, NVIDIA, SickKids, Google, Ubisoft, IBM and other companies, tackled issues focused around the challenges and threats of AI.

The conference kicked off with the “CEOs Talk Tech” opening keynote featuring two CEOs, Sebastien Gendron of TransPod Inc. and Steve Irvine of Integrate.ai. We got to hear about their personal journeys, successes, failures, takeaways and key lessons they learned from their experiences leading them to where they are today. An important message they left with us is to make sure you work with people who share the same big picture and values as you. Everyone wants to be innovative, but how do you find a competitive edge?

“Marry a problem, date a solution…” – Steve Irvine, Founder & CEO, Integrate.ai

The discussions during the two-day conference touched on many key issues that businesses and individuals are facing.

First takeaway: There has been much hype surrounding artificial intelligence, claiming to be the be-all and end-all of all problems. But, in reality AI is not the solution to all our problems or at least we’re not there yet. We should to stay realistic and face the challenges.

An insightful presentation was delivered by Sanja Fidler, where she talked about the advances in AI at NVIDIA. She shared an example of challenges that are facing self-driving cars, which is the image processing technology. A similar point was mentioned by Inmar Givoni from Uber ATG on day 2. Needless to say, deep learning relies on human-labelled data, which is a labour-intensive process before the algorithm starts its training. Tools such as Polygon RNN can effectively assist to label data and reduce human labour.

Sanja Fidler of  NVIDIA, on stage sharing the advances in AI that NVIDIA has made at the AI World Forum hosted at the Carlu in Toronto on March 4-5, 2019.

 

“AI needs to be used to create more labeled data” – Sanja Fidler

During the “Enterprise AI and The Consumer Experience” panel, panelists agreed that business owners needed to be more realistic about how AI can help their business process. Brian Carpenter from Pure Storage also suggested that when looking at your business, it’s critical to understand your data and tackle the problem with specific solutions step by step. The main obstacle for many, is understanding the problem you are facing and the factors that are deterring you from solving it. If you do fail then the strategy is to fail fast and cheap, get real time understanding of what the data is trying to tell you.

Second takeaway: Will AI take our jobs? A daunting question that was raised multiple times throughout the conference.

“We talk too much about replacing jobs, we should also think about augmenting and accelerating jobs” – Chiel Hendriks, Google Cloud

Concerns over job security and replacement by AI has been expressed many times, and the speakers of the robotics panel answered it very well. Think about the 1980’s, when computers were first being introduced to the general public and many people feared of losing their jobs. In reality, the introduction of this new technology helped to create more jobs and work for humans. Instead of fearing AI, we need to utilize it, learn it and create new ways people can contribute to society. Jobs will not be replaced but created, it is up to us to make sure that our skills are aligned with the future because at the end, it is not a rat race against machines, but against ourselves.

“There will be job displacements, but the creation of jobs will be high and possibly more than those that are lost.” – Manir Al-Faisal, Shelley Automation

Third takeaway: How should we treat data correctly and will privacy be a luxury?

Data security and privacy has always been a major area of discussion when it comes to AI.

“We need to ask ourselves what does privacy mean? How should data be collected, stored and shared?” – Matti Siemiatycki, University of Toronto

Privacy is the ability for one to seclude their life from things around them. It is a basic right but with AI, could people be giving up that right without even knowing it? In this information age, our privacy depends solely on how data is stored, modified and exchanged between several parties. A smart toothbrush will record the total time you have brushed your teeth for and then send it to your insurance company. It won’t be just you who knows exactly how long you brushed your teeth.

Our experts from the “Planes, Trains & Automobiles: The Future of Transportation” chimed in with their own privacy concerns with the rise of AI, especially surrounding the area of ethics. People who choose to live in smart cities are giving up their privacy for efficiency. Is there even an option to opt out? Do we get to control or even have a say in what information is being stored and shared?

It was discussed on “The Future of Media and Entertainment” panel that it’s hard to collect authentic data from AI powered VR/AR solutions, as the audience will provide inauthentic reactions because of the fear of cameras around them. Yves Jacquier from Ubisoft Montreal shared the advantages that eSport companies have, which is the asset of authentic animation data in the gaming industry to help close the gap of “wasted ideas”. It’s also been mentioned that collecting data without attributing them to individuals will help protect people’s privacy and train the algorithm with effective data at the same time.

Sarah Siu of Shopify, on stage sharing the role of AI in e-commerce at the AI World Forum hosted at the Carlu in Toronto on March 4-5, 2019.

So, what does AI mean for the future?

Mercedes Benz is coming out with their own chatbot called ‘Ask Mercedes’, which can teach you how to drive as you drive. AI is not just your next Netflix recommendations, we want to see more AI being implemented across all industries but first we need to have the infrastructure in place to support this type of innovation, such as 5G networks. Christine Duhaime, Digital Finance Institute, emphasized that there needs to be a solid infrastructure for 5G as AI continues to develop because it could not only allow a space for remote medical interventions but also a space for collaborative gaming and eSports.

Gaming and eSports expert, Yves Jacquier of Ubisoft, also mentioned there is a gap between gaming and academics that needs to be bridged. He shared an example of how an AI driving game could teach a user how to drive realistically based on a rewards system.

Over the past few years, Shopify has helped its merchants achieve their dreams by providing them with an e-commerce platform. Sarah Siu of Spotify, shared how they have been using augmented reality to allow customers to preview products in real time, before making a purchase. Another way Shopify has been using AI is through their own app store recommendations for first time entrepreneurs. They are able to use AI to help recommend useful apps that it believes can contribute to the success of the business owner.

Another topic that was widely discussed was how AI has helped healthcare professionals better assist their patients. Anna Ansel, proven ag, said that there is a lot of applications for AI in mental health in Europe now because it allows patients to be more open about what exactly they are feeling, allowing doctors to better treat their symptoms. Devin Singh, SickKids, added that before the introduction of AI tools, doctors were more focused on documenting a patient’s symptom rather than connecting with them on a human level, but now they can spend more quality time with their patients and have that human connection. Deborah Raji of Google AI, spoke about the importance of training Google’s Algorithm to be more inclusive. She shared an alarming example to illustrate this glaring problem. When you search for images of “healthy skin”, it was pointed our that the search results rarely included men or women with darker skin. Another example of the lack of inclusion can be found when images of celebrities failed to show or identify famous Asian celebrities. To tackle bias, there needs to be a more diverse dataset that reflects real life demographics and contexts. 

The possibilities of utilizing AI to solve problems are endless, but we cannot let fear stop us from allowing AI and humans to reach their true potential.

What would you do if you weren’t afraid?” – Steve Irvine, CEO of Integrate AI