If you follow the news of Artificial Intelligence, you will get two diversifying threads. Media and cinema often portray AI Human capabilities, Massive unemployment, and a possible robotic apocalypse. On the other hand, scientific conferences discuss progress. Artificial common sense Accepting that Current ai is weak And is incapable of many basic functions of the human mind.
But wherever they stand compared to human intelligence, today’s AI algorithms have already been created A defined component for multiple fields, Including health care, finance, manufacturing, transportation, and many more. And very soon “no field of human endeavor will be independent of artificial intelligence”, as Harvard Business School professors Marco Insity and Karim Lakhani point out in their book Compete in the age of AI: Strategy and leadership when algorithms and networks drive the world.
In fact, weak AI has already led to the growth and success of companies such as Google, Amazon, Microsoft and Facebook, and is affecting the daily lives of billions of people. As Lakhani and Inyanshi discuss in their book, “We don’t need an ideal human replica to prioritize content on social networks, create a perfect cappuccino, analyze customer behavior, set optimal value, or here Until, of course, paint in style. Of Rembrandt. Incomplete, weak AI is already enough to change the nature of companies and how they operate. “
Startups that understand the rules for running AI-driven businesses are able to create new markets and disrupt traditional industries. Established companies that have adapted themselves to the age of AI have survived and thrived. Those who fell into the old ways after losing their ground to companies that exploited the power of AI, either ceased to exist or were marginalized.
One of the many topics Iansiti and Lakhani discuss is AI contacts, the key ingredient that enables companies to compete and grow in the age of AI.
What is AI factory?
key AI technologies are used in today’s business Machine learning algorithms are statistical engines that can brighten patterns from previous observations and predict new results. With other key components such as data sources, experiments and software, Machine learning algorithms AI can build factories, a set of interconnected components and processes that nurture learning and development.
Here’s how the AI factory works. Quality data derived from machine data learning algorithms derived from internal and external sources to make predictions on specific tasks. In some cases, such as diagnosing and treating diseases, they can predict Help human experts in their decisions. In others, such as content recommendation, machine learning algorithms can automate tasks with little or no human intervention.
The AI factory’s algorithm and data-driven model allow organizations to test new hypotheses and make changes that improve their systems. It can be an existing product or new products added to a company that may already have new products manufactured. These changes, in turn, allow the company to obtain new data, improve AI algorithms and re-increase performance, create new services and products, find new ways to grow and move to markets.
“In its essence, the AI factory creates a virtuous cycle between user engagement, data collection, algorithm design, prediction, and improvement,” Iansiti and Lakhani write Compete in the age of AI.
The idea of building, measuring, learning and improving is not new. It has been discussed and practiced by entrepreneurs and startups for many years. But AI factories take this cycle to a new level by entering into such fields natural language processing And Computer vision, Which until a few years ago had very limited software penetration.
Is an example Compete in the age of AI Discussions Ant Financial (now known as Ant Group), a company founded in 2014 with 9,000 employees and providing comprehensive services to over 700 million customers with the help of a very efficient AI factory (and Genius Leadership) Huh. In that perspective, Bank of America, founded in 1924, employs 209,000 people, serving 67 million customers with a more limited array.
Iansiti and Lakhani write, “Ant Financial is a different race.
AI factory infrastructure
It is a known fact that machine learning algorithms rely heavily on data. The value of data has given rise to idioms such as “Data is the new oil,” a cliché has been used. In Several material.
But large amounts of data alone is not good for AI algorithms. In fact, many companies sit on vast reserves of data, but their data and software exist in different silos, stored in inconsistent fashion, and in inconsistent models and frameworks.
“Even though customers see the enterprise as a unified entity, data from systems and units and functions internally are usually fragmented, preventing aggregation of data, delaying insight generation, and analytics and The power of AI is made impossible to take advantage of. ” Ianshi and Lakhani write.
Furthermore, before the AI algorithm is fed, the data must be preprocessed. For example, you can use the history of previous correspondence with customers to develop an AI-driven chatbot that automates parts of your customer support. In this case, the text data must be consolidated, stripped of tokens, excessive words, and punctuation marks, and go through other changes before they can be used to train the machine transformation model.
Even when working with structured data such as sales records, gaps, missing information and other mistakes can occur, which need to be resolved. And if the data comes from different sources, it should be collected in a way that does not cause mistakes. Without preprocessing, you will train your machine learning model on low-quality data, resulting in a poor AI system.
And finally, internal data sources may not be sufficient to develop the AI pipeline. Sometimes, you will need to supplement your information with data obtained from external sources such as social media, stock markets, news sources, and more. An example is Bluedot, a company that uses machine learning. Predicting the spread of infectious diseases. To train and run its AI systems, BlueDot automatically collects information from hundreds of sources, including statements from health organizations, commercial flights, livestock health reports, climate data from satellites, and news reports. Most of the company’s efforts and software are designed to gather and integrate data.
In Compete in the age of AI, The authors introduce the concept of a “data pipeline”, a set of components and processes that consolidate data from various internal and external sources, cleanse data, integrate it, process it, and transform it into various AI systems. However, what is important is that the data pipeline operates in a “systematic, sustainable and scalable manner”. This means that the AI factory must include at least manual effort to avoid causing bottlenecks.
Iyanshi and Lakhani also elaborate on the challenges involved in other aspects of the AI factory, such as setting up the right metrics and facilities Supervised machine learning algorithmsFinding the right division between human expert insights and AI predictions, and tackling the challenges of ongoing experiments and validating the results.
“If data is the fuel that runs the AI factory, then the infrastructure builds the pipes that distribute the fuel, and the algorithms are the machines that work. The experiment platform, in turn, controls the valves that connect new fuels, pipes, and machines to existing operating systems, ”the authors write.
To be an AI company
In many ways, building a successful AI company is an engineering as a product management challenge. In fact, many successful companies have figured out how to build the right culture and processes on long-standing AI technology rather than trying to fit the latest developments Read or learn to meditate In an infrastructure that does not work.
And this applies to both startups and long-standing firms. As Ivansi and Lakhani point out Compete in the age of AI, Technology companies that are left constantly changing their operating and business models.
“For traditional firms, a different kind of organization is about to become a software-based, AI-driven company – those accustomed to ongoing change,” they write. “It is not about closing a new organization, setting up periodical skeletons or creating AI departments. It is about changing the core of the company fundamentally by building a data-centric operating architecture supported by agile organizations, which enables this change. “
Compete in the age of A.I. Is enriched with relevant case studies. It contains stories of startups building AI factories from the ground up like Peeple, which disrupted the traditional home sports equipment market for Ocado, which leveraged AI relying on groceries, a very tight market margin. You will also read about established tech firms as Microsoft, which have managed to thrive in the age of AI through many changes. And there are stories from traditional companies like Walmart, long-time retail giants filed for bankruptcy in 2018, digitization and AI to avoid the fate of the likes of Sears.
The rise of AI has also given new meaning to “network effects”, a phenomenon that has been studied by tech companies since the inception of the first search engines and social networks. Compete in the age of A.I. Different aspects and types of networks and how AI algorithms can be integrated into the network can promote development, learning, and product improvement.
As other experts have already seen, advances in AI will have implications for everyone running the organization, not just those developing technology. Per Iansiti and Lakhani: “Many of the best managers will have to reclaim both the basic knowledge behind AI and learn the ways that technology can be effectively deployed in their organization’s business and operations models. They are not required to become data scientists, statisticians, programmers, or AI engineers; Rather, as every MBA student learns about accounting and its salute to business operations without wanting to become a professional accountant, managers need to do the same with AI and the associated technology and knowledge stack. “
This article was originally published by Ben Dickson TechtalksA publication, which examines technology trends, how they affect the way we live and do business, and the problems they solve. But we also discuss the bad side of technology, the deep implications of new technology, and the things we need. You can read the original article here.
Published January 1, 2021 – 22:00 UTC