In 2016 AlphaGo of DeepMind played a five-game Go match against the 18-time world champion from South Korea, Lee Sadol. AlphaGo won by four against 1. Due to Go’s complexity and the importance of reaction and intuition, it has proved harder for computers to master than simpler games such as checkers or chess. Go has too many moves for a machine to win by brute-force calculations, which is how IBM’s Deep Blue famously beat former world chess champion Garry Kasparov in 1997. DeepMind’s AlphaGo algorithm used a combination of neural networks, machine language, and tree search techniques, combined with extensive training, both human and computer play.
From this game “machine against humans,” four different aspects were derived:
1) AlphaGo initially learned from thousands of games played by professional human players.
2) Using Monte Carlo simulations, AlphaGo decides on the next move based on a probability to win associated with each of the possible activities.
3) AlphaGo improves by playing against itself by using reinforcement learning.
4) AlphaGo uses probabilistic methods, which are based on a representation of the world (a distribution of potential outcomes). As computing power grew exponentially, computers have become better than humans at computation by many orders of magnitude. That said, the game of Go is so complex, and the array of possible moves is so broad that AlphaGo cannot consider all the different possible game situations. When AlphaGo started playing against Lee Sedol, there were still “trajectories” or “game situations” that AlphaGo had not experienced.
Let me comment on only two moves (1-move by AlphaGo and 1-move by Lee Sedol) of the 5-game series.
In game 2, AlphaGo’s move 37 was so surprising they overturned hundreds of years of received wisdom. That is, any average human, who would have tried to maximize the number of points short to mid-term instead of maximizing the probability of winning at the end of the game, would have made a different move. In other words, AlphaGo favored a scenario where it would win by one and a half points with 99% probability over a situation where it would win by 20 points with 80% probability”. And this is because of its outstanding capability to foresee trouble 50 to 60 moves after.
In game 4, Lee Sedol’s move 78 is said to be the cornerstone of his victory. Commentators have called it the “divine move.” The designers of AlphaGO have called it “the one in ten thousand moves” because AlphaGo had calculated that a human would play this move at a probability of one out of ten thousand. Even more interesting is that this move led AlphaGo to make suboptimal decisions in the following rounds. Indeed, AlphaGo’s next ten moves triggered a sharp decrease in its probability of winning the game. It fell from 70% to below 50%, and AlphaGo never managed to go above 50%. Lee Sedol’s strategy during game 4 was to force an “all or nothing” situation instead of trying to gain points in small increments. His idea was that AlphaGo was superior at making the right moves to optimize for small gains thanks to its capability to compute very accurately the probability to win at any point in time but that it could fail in extreme situations where the progress of issues in one move would be more critical.
Computers are, in essence, much better than humans at computing probabilities. Therefore, they can make better decisions about assessing an environment’s different possibilities. But one-way humans beat the machine is to change the setting. In this case, it was about playing such an unexpected move by Lee Sedol that the computer did not even consider it a possibility. This perspective seems to be a good illustration of a situation where creativity defeats computational power. AlphaGo’s defeat in this game shed some light on some of the flaws of Monte Carlo’s methods.
Therefore, intelligent computing platforms can sometimes make better decisions than humans because their computing capabilities allow them to optimize for an outcome that most people cannot predict. But this becomes even more interesting in a context where artificial intelligence does not make decisions but also helps people make the right decisions. This approach is what people call “Intelligence Augmentation.” After Deep Blue (IBM) defeated Gary Kasparov in 1997, people wondered if an AI + Human combination could really beat first a human and then a machine. In 2005, they experimented and found that a Human+AI beats a solo human. Also, they discovered that a Human+AI wins the computer solo. This perspective is because intelligence is not a single dimension, unlike the unscientific IQ tests on the Internet at clickbait sites. The “factor g,” also known as “general intelligence,” represents only 30-50% of a person’s performance in different cognitive tasks. Hence while it is an important dimension, it is not the only dimension. For example, great masters are good at long-term chess strategy but poor at looking ahead to millions of possible moves – and vice versa for AI playing chess. And because humans and AIs are strong on different dimensions, like a colossus, they can beat out solo humans and computers. Integrating AI applications into our day-to-day process could help us improve our decision-making in many different contexts or industries.
Businesses today are looking forward to rapidly adapting to AI technology (e.g., expert systems, natural language processing, speech recognition, and machine vision) in order to survive and stay competitive in today’s highly changing environment because of pandemics, digital transformations, scales of economy, and generational changes. AI applications in healthcare, auto-industry, aerospace, social media, video games, travel navigation, banking & finance, smart home devices, security and surveillance, and e-commerce are enhancing and becoming part of our lives daily. AI scholars and game-changers believe that the future of consumer behavior by 2030 will be digital, mobile, and deeply Internet engaging.
In B2C, today is required to prepare and overcome several challenges before it can fully realize the benefits of high digital technology, such as selecting the right AI platform, Changing company culture, Processing vast amounts of data, and becoming very agile. Effective AI requires two crucial steps: (1) Plan your AI migration carefully; (2) Establish a pace to prevent holes in data. All of the above requires precise AI analytics and data. To achieve a transformation strategy, one needs to ask, when developing their plan, the following: (1) Where is data being held across the marketing stack? (2) Where is data transferred across the marketing stack? (3) What marketing efforts will be affected by data loss? A marketing technologist must use data to help his teams draft better campaigns and become more creative; Analyse software capabilities within the needs of his organization; Function as the hybrid of marketing and IT; Break out of legacy organizational systems as needed. Yet, there appears to be a severe imbalance between consumer technology and marketing technology (i.e., a set of software solutions used by business leaders to support mission-critical business objectives and drive innovation within their organizations) since it is tough to gain a leading edge that is constantly advancing. Thus, businesses today require new capabilities to close the consumer and marketing technology gap. To do so, there must be a holistic database and technology strategy that they can individualize at scale, customer journey capabilities that can adapt in real-time, and intelligence to automate the self-reinforcing cycle of tailored experiences. Tomorrow’s customer journey and personalization will be even more brilliant, immersive, and trust-enabling. More customer experience initiatives will be run by AI and machine learning algorithms, automated software applications, or bots. In this case, the question is whether the brands and consumers will be ready.
Let me add, however, that today B2C artificial intelligence applications create misleading statements behind the data, while ethics lag in terms of their comprehensive customer service. From a cynical view, I believe that most companies are not interested in the social contract. At the beginning of the 21st century, ethics and security issues are becoming some of the most profound challenges to each AI business strategy, especially when 75% of the public is highly concerned with how companies use their personal data. And although companies report that they are embracing technologies like facial recognition software, consumers say that they do not like these technologies. Of course, these beliefs directly result from the customer’s negative experiences and their discomfort with the lack of control they feel they have over their privacy and data. As a result, I’m outraged by how different business brands are taking advantage of the masses’ data today while simultaneously creating a profound distrust and inconsistency of performance over customers’ data. In this regard, I believe that the leaders of these companies, in the name of profit and corporate ambition, fail to be the real game-changers because they fail to create clarity, be truly innovative for the common good, think outside the box, release energy friendly to the customer, compromise with anything more than ordinary, build trust and show no empathy for others. Thus, it appears to me that a lot of companies and brand experiences show that they don’t share value, but they are exploitative.
Although consumer preferences for technology are focused on their experience from the perspective of users, businesses see 2030 as an opportunity to leverage integrated AI technology combined with advanced customer behavior to provide the intelligence behind engagement, leading to deeper, more meaningful customer relationships and increased loyalty. Specifically, companies recognize that innovative home systems will have the intelligence and ease of access to be not only a key gateway but also a gateway between business and customer and that artificial intelligence, machine learning, and predictive analytics items can help companies to get involved and a more comprehensive view of the customer. Thus, this will allow the businesses to deliver on their loyalty promise beyond the fundamentals and be a business that inspires loyalty, not only a business offering.
The B2C relationship was once exclusively human-to-human, but by 2030 it is believed that it will be primarily driven by machine-to-human and then by machine-to-machine facilitated by humans. I will not judge if this is a good or bad thing. Instead, I will state that businesses must ensure that the elements of trust are present before and throughout the relationship and that the gaps between company and customer are addressed today and proactively in the future, and to achieve that is necessary to develop a further complete and accurate 3600 understanding of the customer with the resources and value needed to inspire and create trust in the businesses’ brands.