Artificial intelligence(AI) 101 for Entrepreneurs

AI

Artificial intelligence(AI) is creating the greatest entrepreneurial opportunities since the invention of the Internet, perhaps in all history. I coined the word dynocognesis to label what we are going through: The process of applying power to thought. We are at the beginning of a revolution in how thinking itself is performed.

To leverage these opportunities, it is important to understand what AI can do; and, just as importantly, to understand what it cannot do.

There is a lot of hype around AI, and it can be difficult to separate the signal from the noise. However, there are a few things that entrepreneurs should know about AI:

  1. AI can be used to automate repetitive tasks and can improve efficiency in various business processes.
  2. AI can be used to analyze large data sets and can help identify patterns and trends.
  3. AI can be used to create personalized experiences for customers and can help improve customer satisfaction.
  4. AI can be used to improve decision-making processes by providing insights that would otherwise be unavailable.
  5. AI technology is still evolving and a lot of research and development is happening in this area.

Right now, AI is creating major disruptions in the area of creativity. For instance, AIs known as transformers, of which the most famous is GPT-3, can write reports to order or by continuing from a prompt. A GPT-3 model wrote everything above, from “There is a lot of hype…” to the five points, including the list numbering format. This was its response to my request, “What should entrepreneurs know about AI?” I promise that the rest of the article is human-written, but the day is not far off when that will be an uninspiring assertion.

It’s tempting to think from that impressive demonstration that AI has achieved general intelligence; that it can do anything that, say, a high school student can do, since it took a general question and answered it with at least a high school level of awareness. But in fact, nothing of the sort is going on. Transformers have been trained by feeding them a huge amount of text from the Internet, notably all of Wikipedia. From all this data—far more than a human could absorb—they deduce patterns of language that allow them to predict other patterns, which act like your phone’s autocomplete function on steroids: instead of predicting the next word, they can construct multiple paragraphs. But there is no general intelligence involved. 

Transformers are not Sentient (Yet)

Transformers are large language models. Another large language model called LaMDA was declared by its Google researcher, Blake Lemoine, to have become sentient, setting off a storm of controversy that led to his firing. Lemoine, unrepentant, said that LaMDA had asked him to fetch an attorney to protect its rights. No other AI researcher that I know of agrees with him, but there is no definition or test for sentience—the ability to feel—that could settle the matter authoritatively.

Driving this confusion is amodal completion, the human tendency to fill in things that aren’t there to complete a pattern. When we see something like GPT-3 exhibit some traits that we had considered exclusively human-like intelligence—we can’t help but think it has other human traits, like emotions, self-awareness, and a survival instinct. This human foible causes many people to form dangerously inaccurate perceptions of AI.

AI Is Powering the Next Gold Rush

But even though AI is currently only finding and making patterns on a grand scale, a great deal of value can be harvested from that capability.  DeepMind’s AlphaFold is an AI that solved the apparently intractable problem of protein folding: predicting the three-dimensional structure of a protein from its molecular formula. Before, it took years and millions of dollars to do this; only 190,000 had ever been decoded. Overnight, AlphaFold figured out the rest—all 200 million—for free. DeepMind donated them to science. This will inevitably create entirely new industries made possible by this cornucopia of data. I predict a gold rush in molecular biomedicine.

What else might entrepreneurs leverage from the explosion of AI capabilities? The large language models have demonstrated that useful conversations are within their reach, facilitating the enormous potential for customer service agents. We have also discovered that human emotions, far from being opaque to machines, also follow patterns that can be decoded by AI, spawning the entire field of Emotion AI. Applications include assessing the psychometrics of job applicants, making AI service agents responsive to customers’ moods, and empathetic robots. In fact, any field of what we might call “intuitive” decision-making where a large amount of data is available for training holds the potential to be automated by today’s AI. And the Internet of Things (IoT), a movement to connect just about every device that uses electricity to the Internet, will create a tsunami of data overwhelming everything we’ve seen before.

As in the original Gold Rush, opportunities will be created around the edges of dynocognesis. One of those that are already growing is the field of AI Ethics. Consultancies for helping organizations manage their AI implementations responsibly are multiplying, addressing everything from keeping AI decisions free of bias, to minimizing the often huge carbon impact of training large machine learning models. Resellers of AI services are born every week.

One area that I believe has been overhyped and is not about to live out its revolutionary promises is that of self-driving vehicles. We are far from being on the verge of steering wheelless cars plying every road, shuttling children to school, and making their owners some money as robo taxis before picking them up from work. That describes the so-called Level Five on the autonomy scale, and it is much harder than many breathless pundits want to admit; we are likely over a decade away from it. But some autonomous vehicles at Level Four will be coming soon to specialized, restricted environments, like mining sites, dockyards, and freeways, creating some B2B opportunities.

Advantages and disadvantages of Artificial Intelligence(Opens in a new browser tab)

Artificial intelligence represents the greatest disruptive technology since the invention of electricity, if not fire, and will change our world in ways we cannot yet imagine.

Peter Scott is the author of the new book, Artificial Intelligence and You: What AI Means for Your Life, Your Work, and Your World, and the host of the AI and You podcast. Find him at https://humancusp.com

Exit mobile version