Big Data and AI Toronto 2019 – Can we trust AI?

Attending a Big Data Toronto annual events in mid-June became my tradition since the first Big Data Toronto conference in 2016. (1,2,3) For me, the event is a great opportunity to get the feel of the broad progress in the area of Big Data and AI year over the year. I believe that Big Data and AI 2019 (4) was the most exciting event so far.

The conference organizers invited speakers that represented a variety of businesses and government organizations who also play different roles within their respective entities. We were able to hear a broad range of perspectives about AI technology’s impact on specific organizations and society as a whole.

AI hype vs. reality

You may hear and read in the media different opinions about AI:

• Skeptical: AI is overhyped, and there is not much practical use for most people.
• Excitement: AI technologies are driving the next industrial revolution and will cause significant changes within the next 5 – 15 years. We need to ride the wave. Technology adoption is accelerating, and we will adopt AI technology in our daily life with the same velocity as smartphones.
• Conservative: we are only at the beginning of the AI revolution, and as with the adoption of electricity, it will take decades for the gradual transformation of business and society.
• Fatalist: AI is more dangerous than nuclear weapons and may lead to the end of human civilization.

The reason for such a wide range of opinions is that the majority of people have difficulties in understanding and explaining AI.  Even experts use the term Artificial Intelligence in multiple contexts.

“Only 4 percent of Canadians felt confident about explaining what AI is.”

Alex Grbic made an excellent presentation based on Canada’s AI imperative research published by Deloitte (5). Here are some facts that illustrate the state of AI understanding and adoption in Canada:

• Only four percent of Canadians surveyed feel confident enough to explain what AI is and how it works.
• Eighty-six percent of Canadians said they don’t use AI-powered tools or devices; given that 76 percent of Canadians use a smartphone that has virtual assistant programs and mapping software, there is a disconnect between what Canadians know about AI and the reality of AI technology adoption.

Many people do not realize that AI technologies have been around for decades. For instance, I worked with some of the neural network applications already in 1991. Since the nineties, AI technologies were experiencing steady progress due to the increase in computer power and the development of new algorithms.

The rapid adoption of the Internet brought new business applications of machine learning technologies in e-commerce and digital advertising. Predictive modeling, natural language processing, pattern and image recognition have been embedded in multiple applications. However, media and public did not raise much concern about that progress. Most people, until relatively recently considered Artificial Intelligence as the subject of sci-fi movies or books.

The attitude towards AI suddenly changed when the public became aware that many technologies that are foundational elements of AI are already part of their daily life. You can not deny that AI technologies are here today when you are listening to the conversations between kids and Alexa.

Millions of consumers experienced AI technologies embedded in electronics devices or customer support systems powered by natural language processing.

The speed of AI technologies introduction in consumer market created perceptions that the development and adoption of AI technologies are progressing extremely fast. Such perception attracted media attention and raised the level of public anxiety about the future of society ruled by AI.

What are the risks of AI? Can we Trust AI?

AI needs more than a technical fix—it requires building trust.

Canada’s AI imperative study (5) provides the following findings of business and consumers attitudes towards AI:

• Ethical risks of AI are top of mind for both businesses and consumers.
• Canadian companies’ reservations about AI often involved AI’s “black box” problem.
• Among Canadian consumers, 65 percent reported privacy concerns over how companies were using their data today
• A few conference presenters touched the AI risks and trust issues as one of the critical obstacles for AI adoption.

The Deloitte report attempts to classify risks into three categories:

Societal risks

The way that AI interacts with citizens can pose risks from a societal perspective. Consider, for instance, the use of automated surveillance which could infringe on individual and cultural views of privacy and autonomy, or the use of automated drones for military warfare.

Economic risks

The transformational nature of AI can cause significant changes to current economic and market structures. For instance, the changing nature of jobs and the workforce can potentially alter employment opportunities for individuals and organizations.

Enterprise risks

Enterprise risks arise as an organization looks to incorporate AI into its strategic and technical frameworks. These risks might include:

• Privacy: AI creates the potential for personal data to be misused or compromised.
• Transparency: The use of black-box models can limit the ability to explain decisions or audit outcomes.
• Control: AI creates the potential for AI systems or models to produce unintended consequences.
• Cybersecurity: AI systems, models, and data may be vulnerable to cyber threats and adversarial attacks.

One of the critical conclusions in the Deloitte report is that AI needs more than a technical fix—it requires building trust.

To take full advantage of AI’s potential benefits, we need to build public trust—something sorely missing today. Canadians’ feelings of distrust in AI go deep; they are rooted in emotion and can’t be overcome by minor technical fixes.

Distrust in AI is the reason that the gap between supply and demand in the Canadian business market is growing. While there is visible growth of companies that offer new products and services using AI technologies, the AI adoption among Canadian businesses is stagnant.

Only 16 percent of Canadian companies use AI, which remains unchanged since 2014

The outlook on AI adoption among Candian business does not look strongly optimistic. The study shows that only eight percent of Canadian companies surveyed plan to increase their spending by more than 20 percent in the upcoming year – 40 percent fewer than the global average.

This fact may have very negative consequences for the future of the Canadian economy. Technology adoption and productivity growth are closely related. Unless the technology adoption accelerates the competitiveness of the Canadian economy will continue to degrade.

The main challenge isn’t technology. It’s culture.

Many people compare the impacts of AI adoption on the economy to the adoption of electricity during the second industrial revolution. I would agree that the impact of AI adoption in society is going to be even more significant than electricity. The use of electricity fundamentally changed the way people worked and lived.

AI will change not only the way how people work and live but all elements of society:

• Government
• Education
• Juridical system
• Healthcare
• Media
• Industrial and business processes
• Culture

We see examples of how business and government use AI-powered technology to benefit consumers, patients, students or constituents.

AI and e-Governance

“The aim is to make government more proactive and responsive to people’s life-events”.

If somebody asks me what the most notable presentation from the Big Data and AI 2019 agenda was, then my answer would be “Taking the Next Steps in E-governance” by Ott Velsberg who is Chief Data Officer Government of Estonia.

Estonia is a small nation 1.3 million people that gained back its independence after the collapse of the Soviet Union.

Estonia has become one of the world’s most developed digital societies. Estonian government implemented digital ID for all citizens since 2002, online voting since 2005, and put digitized health records on the blockchain more than a decade ago.

Ott Velsberg toked provide examples about 16 government processes that have been automated using AI and many more that are now under development or testing.

Here are some impressive examples from his presentation:

• Around 72% of people who were suggested a job by the AI system were still employed six months later, compared with 58% of those advised by a human.
• Estonian Ministry is overseeing the development of a “robot judge” that could decide small claims disputes. The project is in its early phases and will likely start later this year with a pilot focusing on contract disputes.
• Estonia will automate the process of enrolling children in school by the end of this year.   Parents will not need to enroll their children in school, instead, the system will calculate a child’s school needs based on birth data obtained from hospitals.

The foundation of e-government success is the trust between Estonian citizens and their government. The government attention to data privacy was playing an essential role in building the trust between government and citizens. Estonian government applied blockchain technology to protect citizens data.

Ott Velsberg said, “We’re not telling people what to do, we are giving options to consider. You can still make up your mind. The government can never go to the individual level of telling you what to do. That might happen in China but not in Estonia, or Europe as a whole.”

Conclusion

Many presentations during the Big Data and AI 2019 event were about the non-technological aspects of BigData and AI. Technological challenges are still there, but further progress is really about economic and cultural shifts brought by AI technologies. The AI revolution is here today and will continue expanding.

There are no facets of human work and life that will not be changed. Ride the wave, embrace change, and don’t be left behind.