The latest technology and digital news on the web

Human-centric AI news and analysis

What the hell is an AI factory?

If you follow the news on bogus intelligence, you’ll find two deviating threads. The media and cinema often portray AI with human-like capabilities, mass unemployment, and a accessible robot apocalypse. Scientific conferences, on the other hand, altercate advance toward artificial accepted intelligence while acknowledging that current AI is weak and butterfingers of many of the basic functions of the human mind.

But behindhand of where they stand in allegory to human intelligence, today’s AI algorithms have already become a defining basic for many sectors, including health care, finance, manufacturing, transportation, and many more. And very soon “no field of human endeavor will remain absolute of bogus intelligence,” as Harvard Business School advisers Marco Iansiti and Karim Lakhani explain in their book .

In fact, weak AI has already led the growth and success of companies such as Google, Amazon, Microsoft, and Facebook, and is impacting the daily lives of billions of people. As Lakhani and Iansiti altercate in their book, “We don’t need a absolute human replica to accent agreeable on a social network, make a absolute cappuccino, assay chump behavior, set the optimal price, or even, apparently, paint in the style of Rembrandt. Imperfect, weak AI is already enough to transform the nature of firms and how they operate.”

Startups that accept the rules of active AI-powered businesses have been able to create new markets and agitate acceptable industries. Acclimatized companies that have acclimatized themselves to the age of AI survived and thrived. Those that stuck to old methods have ceased to exist or become marginalized after losing ground to companies that have acclimatized the power of AI.

Among the many topics Iansiti and Lakhani altercate is the abstraction AI factories, the key basic that enables companies to attempt and grow in the age of AI.

What is the AI factory?

competing in the age of ai book cover
Competing in the Age of AI by Marco Iansiti and Karim Lakhani

The key AI technologies used in today’s business are apparatus acquirements algorithms, statistical engines that can glean patterns from past observations and adumbrate new outcomes. Along with other key apparatus such as data sources, experiments, and software, machine acquirements algorithms can create AI factories, a set of commutual apparatus and processes that breeding acquirements and growth.

Here’s how the AI branch works. Quality data acquired from centralized and alien sources train apparatus acquirements algorithms to make predictions on specific tasks. In some cases, such as analysis and analysis of diseases, these predictions can help human experts in their decisions. In others, such as agreeable recommendation, apparatus acquirements algorithms can automate tasks with little or no human intervention.

The algorithm– and data-driven model of the AI branch allows organizations to test new hypotheses and make changes that advance their system. This could be new appearance added to an absolute artefact or new articles built on top of what the accession already owns. These changes in turn allow the accession to obtain new data, advance AI algorithms, and again find new ways to access performance, create new casework and product, grow, and move across markets.

“In its essence, the AI branch creates a blameless cycle amid user engagement, data collection, algorithm design, prediction, and improvement,” Iansiti and Lakhani write in .

The idea of building, measuring, learning, and convalescent is not new. It has been discussed and accomplished by entrepreneurs and startups for many years. But AI factories take this cycle to a new level by entering fields such as natural accent processing and computer vision, which had very bound software assimilation until a few years ago.

One of the examples  discusses is Ant Banking (now known as Ant Group), a accession founded in 2014 that has 9,000 advisers and provides a broad range of banking casework to more than 700 actor barter with the help of a very able AI branch (and genius leadership). To put that in perspective, Bank of America, founded in 1924, employs 209,000 people to serve 67 actor barter with a more bound array of offerings.

“Ant Banking is just a altered breed,” Iansiti and Lakhani write.

The basement of the AI factory

artificial intelligence
Image credit: Depositphotos

It is a known fact that apparatus acquirements algorithms rely heavily on mass amounts of data. The value of data has given rise to idioms such as “data is the new oil,” a cliché that has been used in many articles.

But large volumes of data alone do not make for good AI algorithms. In fact, many companies sit on vast stores of data, but their data and software exist in abstracted silos, stored in an inconsistent fashion, and in adverse models and frameworks.

“Even though barter view the action as a unified entity, internally the systems and data across units and functions are about fragmented, thereby preventing the accession of data, dabbling acumen generation, and making it absurd to advantage the power of analytics and AI,” Iansiti and Lakhani write.

Furthermore, before being fed to AI algorithms, data must be preprocessed. For instance, you might want to use the history of past accord with audience to advance an AI-powered chatbot that automates parts of your chump support. In this case, the text data must be consolidated, tokenized, bare of boundless words and punctuations, and go through other transformations before it can be used to train the apparatus acquirements model.

Even when ambidextrous with structured data such as sales records, there might be gaps, missing information, and other inaccuracies that need to be resolved. And if the data comes from assorted sources, it needs to be aggregated in a way that doesn’t cause inaccuracies. After preprocessing, you’ll be training your apparatus acquirements models on low-quality data, which will result in AI systems that accomplish poorly.

And finally, centralized data sources might not be enough to advance the AI pipeline. Sometimes, you’ll need to accompaniment your advice with alien sources such as data acquired from social media, stock market, news sources, and more. An archetype is BlueDot, a accession that uses apparatus learning to adumbrate the spread of communicable diseases. To train and run its AI system, BlueDot automatically gathers advice from hundreds of sources, including statements from health organizations, bartering flights, livestock health reports, altitude data from satellites, and news reports. Much of the company’s efforts and software is advised for the acquisition and accumulation the data.

In , the authors acquaint the abstraction of the “data pipeline,” a set of apparatus and processes that consolidate data from assorted centralized and alien sources, clean the data, accommodate it, processes it, and store it for use in altered AI systems. What’s important, however, is that the data activity works in a “systematic, sustainable, and scalable way.” This means that there should be the least amount of manual effort complex to avoid causing a aqueduct in the AI factory.

Iansiti and Lakhani also expand on the challenges complex in the other aspects of the AI factory, such as establishing the right metrics and appearance for supervised apparatus acquirements algorithms, award the right split amid human expert acumen and AI predictions, and arrest the challenges of active abstracts and acceptance the results.

“If the data is the fuel that powers the AI factory, then basement makes up the pipes that bear the fuel, and the algorithms are the machines that do the work. The analysis platform, in turn, controls the valves that affix new fuel, pipes, and machines to absolute operational systems,” the authors write.

Becoming an AI company

data charts

In many ways, architectonics a acknowledged AI accession is as much a artefact administration claiming as an engineering one. In fact, many acknowledged companies have ample out how to build the right ability and processes on long-existing AI technology instead of trying to fit the latest developments in deep learning into an basement that doesn’t work.

And this applies to both startups and abiding firms. As Iansiti and Lakhani explain in , technology companies that survive are those that continuously transform their operating and business models.

“For acceptable firms, acceptable a software-based, AI-driven accession is about acceptable a altered kind of organization—one acclimatized to advancing transformation,” they write. “This is not about spinning off a new organization, ambience up the casual skunkworks, or creating an AI department. It is about fundamentally alteration the core of the accession by architectonics a data-centric operating architectonics accurate by an agile alignment that enables advancing change.”

 is rich with accordant case studies. This includes the belief of startups that have built AI factories from the ground up such as Peleton, which disrupted the acceptable home sports accessories market, to Ocado, which leveraged AI to digitize groceries, a market that relies on very tight profit margins. You’ll also read about acclimatized tech firms, such as Microsoft, that have managed to thrive in the age of AI by going through assorted transformations. And there are belief of acceptable companies like Walmart have leveraged digitization and AI to avoid the fate of the likes of Sears, the longstanding retail giant that filed for defalcation in 2018.

The rise of AI has also brought new acceptation to “network effects,” a abnormality that has been advised by tech companies since the founding of the first search engines and social networks. discusses the assorted aspects and types of networks and how AI algorithms chip into networks can boost growth, learning, and artefact improvement.

As other experts have already observed, advances in AI will have implications for anybody active an organization, not just the people developing the technology. Per Iansiti and Lakhani: “Many of the best managers will have to retool and learn both the basal ability behind AI and the ways that technology can be finer deployed in their organization’s business and operation models. They do not need to become data scientists, statisticians, programmers, or AI engineers; rather, just as every MBA apprentice learns about accounting and its appendage to business operations after absent to become a able accountant, managers need to do the same with AI and the accompanying technology and ability stack.”

This commodity was originally appear by Ben Dickson on TechTalks, a advertisement that examines trends in technology, how they affect the way we live and do business, and the problems they solve. But we also altercate the evil side of technology, the darker implications of new tech and what we need to look out for. You can read the aboriginal commodity here.

Appear January 1, 2021 — 22:00 UTC

Hottest related news