Big Data Toronto 2018 Conference and Expo (1) was the third event that I attended since the conference inception in 2016.
I must admit that the show made big progress since the first BigData Toronto in 2016. During the first conference, many presentations were about traditional Business intelligence and analytics rather than real Big Data. The title of my blog “Big Data Is Still Like Teenage Sex” was the reflection of my impression from that event. Last year BigData conference was extended to include topics about artificial intelligence and The Internet of Things (IoT).
Here are brief facts about the 3 past events:
| Year | Conference Type | Attendees | Speakers | Sponsors &exhibitors |
|---|---|---|---|---|
| 2016 | One show - Big Data | 1000+ | 50+ | 40+ |
| 2017 | 3-in-1 show Big Data, AI and IoT | 4500 | 120 | 75+ |
| 2018 | 2 collocated shows Big Data and AI | 5000 | 150+ | 80+ |
The 2018 event had only two collocated conferences BigData and AI and did not include IoT as a parallel stream. I think that the organizers made the right decision with the more focused show and topics choices. The conference organizers selected sponsors and speakers to represent a mix of the new generation of AI and Big Data companies as well as familiar names like Amazon, IBM, and Microsoft.
Below are my key takeaways from Big Data Toronto 2018 Conference & Expo.
Change Everything with AI
Change Everything with AI was the presentation by Leon Katsnelson, CTO and Director of Emerging Technologies at IBM. I think that this title reflects the dominant theme of the BigData 2018 conference and Expo. Conference speakers provided many examples of how AI technologies are finding ways into various applications almost in every aspect of human life as finance, healthcare, marketing, government, education, manufacturing, and design.
Progress in AI raises more questions than answers
We recognize many new open questions about how the AI’s disruptive nature may bring unintended and unforeseen consequences. Presenters talked about the influence of AI on decision making, security, privacy, legal liability, education, unemployment and nature of work.
Decision Making and Cognitive Bias
When we are applying AI algorithms to improve decision-making process, we assume that AI helps to eliminate or reduce cognitive biases that unavoidably affect human decisions. Few presenters spoke about “bad data-in- bad-decisions-out” predicament. If we use dirty or biased data to train AI for automated decision-making, we make the quality of decisions rather worse than better. It is more difficult for us to recognize a hidden bias in data used for AI training than in traditional analytics. Humans who prepare training data for AI decision-making systems may introduce biased data unintentionally or intentionally. Intentionally embedded bias in the AI training is not easy to recognize. Criminal or rogue state agents may try to exploit such vulnerabilities and “corrupt” AI to achieve their illegal or immoral goals.
AI and Privacy
Marketers started to use AI to influence consumer behavior since early days of e-commerce. We know for a long time about algorithms used for product recommendations and online advertising targeting. The growth of social media provides more behavior data and new opportunities for AI-based targeted advertising. Political campaign managers also started to use Big Data AI algorithms to influence elections. The well-known first example of using social media in a political campaign was the presidential election in 2008. Obama’s 2008 campaign used not only Facebook and YouTube but also MySpace, Twitter, Flickr, Digg, BlackPlanet, LinkedIn, AsianAve, MiGente, Glee, and others (2). Media was welcoming an innovative approach to the presidential campaign during that time. In contrast, we see entirely different attitude towards using Big Data, AI, and social media for political purposes in 2018. We are much more concerned about abuse of privacy by politicians, marketers and some cases foreign states agents to influence people behavior.
Privacy issues were a big part of the panel discussions about AI regulations and also the topic of the keynote speech by Pamela Snively, Chief Data and Trust officer at Telus.
The privacy issues have many unanswered questions.
It seems that the privacy problem is an example where available technology capabilities are much more ahead from our society’s ability to recognize and deal adequately with the challenge.
AI and regulations
AI and Regulations was another relatively new topic of discussion during the conference. How our legal system has to deal with the new reality where self-driving cars, robots as financial or legal advisors, AI in healthcare become a norm? One of the panel discussions titles was “Looking Ahead at AI Regulations.” Majority of participants were in agreement that AI advances will require new laws and regulations, but there is a high-level of the uncertainty in balancing the government control and innovation.
AI and Unemployment
Will AI create mass unemployment? I heard this question from more than one of the keynote speakers, and it was also mentioned in one of the panel discussion.
One of the first industries experienced a change in its workforce was finance. Speakers mentioned the story about Goldman Sach replacing 600 traders with two traders. Automated trading programs have taken over the rest of the work, supported by 200 computer engineers (3).
Just a few days after the conference Manulife announced about cutting 700 jobs as it digitizes and automates its customer service operations (4).
It is difficult to predict what type of jobs will be safe and not replaced by AI. We know current examples of how AI is applied to make a robotic version of financial and legal advisors, medical diagnostic experts, vehicles drivers, logo or fashion designers, and event music composers. These are relatively new areas in comparison to already well-known examples of robots in the manufacturing processes.
Most of the speakers who touched the topic of AI and employment emphasized that nature of work will change with the evolution of AI. The big question that does not have an answer yet is how workforce and society will adapt to accelerated changes that development in AI may bring in 5-10 years and beyond? We anticipate that employee retraining and continues education will become an integral part of employment lifecycle. I heard a few times during the discussions about an idea to use tax profits from AI based modernization to fund employee retraining.
AI and Retail
I did not see any significant participation of the retail sector at the conference. Few presenters provided examples of current AI applications in retail. Here are just some of the
- Inventory Optimization
- Customer Relationship Prediction
- Demand Forecasting
- Dynamic Price Optimization
- Retail Customer Churn
The benchmark of utilizing AI in retail is Amazon Go physical store that works without cashiers and checkout lines. Amazon applied computer vision and machine learning to create a store where customers could simply take what they want and go. (5)
Conclusions
I touched here only a small number of topics that the attendants could learn during the conference. My main learning was how organizations are pushing boundaries of BigData and AI to create new ways of generating value to consumer and society.
References
- BIG DATA TORONTO 2018 CONFERENCE & EXPO, https://www.bigdata-toronto.com/
- Barack Obama and the Facebook Election, https://www.usnews.com/opinion/articles/2008/11/19/barack-obama-and-the-facebook-election
- As Goldman Embraces Automation, Even the Masters of the Universe Are Threatened, https://www.technologyreview.com/s/603431/as-goldman-embraces-automation-even-the-masters-of-the-universe-are-threatened/
- Manulife cutting 700 jobs, vacating office in downtown Kitchener, https://www.therecord.com/news-story/8686505-manulife-cutting-700-jobs-vacating-office-in-downtown-kitchener/
- Amazon Go, https://www.amazon.com/b?ie=UTF8&node=16008589011
