AI wrote half this article, can you tell?

The future is already here — it’s just not evenly distributed.” ― William Gibson

By 2025 half of all workplace tasks will be carried out by a computer according to the World Economic Forum. It isn’t necessarily that people will be replaced but our work will be augmented in new ways through artificial intelligence. I really wanted to understand what that would look like and feel like, so I’m writing this article with the assistance of artificial intelligence. First, some background.

My interest in AI led me to learn about OpenAI. OpenAI was founded in 2016 by Tesla’s former CEO, Elon Musk and other notable tech figures. Their goal is to provide a platform and offer software tools for artificial intelligence (AI). These tools can be used by researchers, programmers, companies or everyday people who want to study machine learning and AI without having to spend years programming it themselves. The company has grown since its inception and now employs around 50 people with PhDs from various fields of expertise.

OpenAI’s primary objective is to build safe and beneficial AI for the world.

What OpenAI does is develop general-purpose algorithms that are not constrained by specific inputs or tasks, but instead can be applied like a human expert would do to any unknown problem (similar to Alpha Go). The hope is that these general-purpose algorithms will allow humans to make better use of the time they have by making them less dependent on specific skillsets. One of these general-purpose algorithms is GPT-3.

GPT-3 is a general purpose language model that can generate coherent paragraphs of text with just one line of input. It is a text generator that is able to produce coherent, grammatical sentences. The algorithm can go from one sentence to the next without interruption and it has a variety of sentence structures. For example, it can produce sentences that are compound-complex, such as: “John ate all the donuts and then he brushed his teeth.”

GPT-3 is not an application itself, you need to access it through an API. It turns out a number of content writers are built on top of GPT-3 including Jasper.AI, Rytr.AI and Copy.AI. I used Rytr.AI for this article. By providing a topic sentence and some keywords, it then generates a paragraph. Most of the generated text in this article I didn’t change. I copied and pasted it into this article. How much for this magic tool? Free trials are available and full versions are available for as little as $9/month.

Much of what we experience of AI today is pattern recognition. AI trains on a data set and then when given a test item identifies it. This shows up from facial recognition through to speech to text. This article is an example of the next evolution of AI, Generative Artificial Intelligence (GAI). GAI is a new intelligence system, which was specifically developed to create content. It is supposed to create content that is worth publishing and engaging.

Much of the products using generative text are focused on the marketing industry to assist in areas such as creating headlines, product pitches and blog content. A study that was published in Creative AI magazine showed how a creative workflow driven by artificial intelligence could produce over 6,000 headlines in an hour. Imagine new areas where this could apply such as generative news. Based on your personal preferences and history, not only would news articles relevant to you be surfaced but the content itself would be written specifically for you.

It’s not in 2025 that we have AI-assisted work, it is happening now. As William Gibson would say it’s just not evenly distributed. Today the generative feature exists in niche, start-up software products. Consider though that Microsoft, whose office productivity software is ubiquitous, invested $1 Billion in acquiring the use of GPT-3, and it would only make sense these features eventually show up when we start up Word. Those people and companies who start leveraging these early AI-assisted tools today will start to exceed the productivity of their peers ahead of when we hit that mainstream number of 50% of workplace tasks conducted by computer in 2025.

Originally published at




Lifelong exploration and teaching of Leadership, Technology and Finance.

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

ML Engineers Are Losing Their Jobs. Learn ML anyway

UTU and Curacel Partner to Enhance Trust in Insurance Provision across Africa

UTU and Curacel partner for socially-powered insurance provision

Synthetic Abstractions

Empathy in AI Series: Part 2, Empathetic Machines Creating Jobs?

Linear Regression v.s. Neural Networks

Advancements in Semi-Supervised Learning with Unsupervised Data Augmentation

Validate IQBot OCR engine Results

AI in government: for whom, by whom?

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Craig McQueen

Craig McQueen

Lifelong exploration and teaching of Leadership, Technology and Finance.

More from Medium

Experiment: Trying out Houdini 19 Vellum, Vellum Fluid and Interactions with Each Other

Interview With J Stewart Willis, Author of “Deadly Highway”

CS373 Spring 2022: Ziyi Zhao

Good news: I finally made it to grad school!