One of the biggest trends in machine learning right now is text generation. AI systems learn by absorbing billions of words scraped from the internet and generating text in response to a variety of prompts. It sounds simple, but these machines can be used for a wide variety of tasks – from creating fiction to writing bad code and letting you chat with historical figures.
The best-known AI text generator is OpenAI’s GPT-3, which the company uses recently announced is now used in over 300 different apps, by “tens of thousands” of developers, and produces 4.5 billion words per day. That is one a lot of of robot verbiage This may be an arbitrary milestone for OpenAI to celebrate, but it is also a useful indicator of the growing scale, impact, and commercial potential of AI text generation.
OpenAI started life as a non-profit organization, but in recent years it has been trying to make money with GPT-3 as the first marketable product. The company has an exclusivity agreement with Microsoft that gives the tech giant unique access to the program’s underlying code, but any company can request access to GPT-3’s common API and build services on top of it.
Because OpenAI likes to advertise, hundreds of companies are now doing exactly this. A startup named Profitable uses GPT-3 to analyze customer feedback and identify “themes, emotions and sentiment from surveys, help desk tickets, live chat logs, reviews, and more”; Fable Studio uses the program to create dialogue for VR experiences; and Algolia uses it to improve its web search products, which it in turn sells to other customers.
This is all good news for OpenAI (and Microsoft, whose Azure cloud computing platform powers OpenAI’s technology), but not everyone in Startup-Land is enthusiastic. A lot of analysts have commented that it is foolish to build a business on technology that you don’t actually own. Using GPT-3 to create a startup is ridiculously easy, but it will also be ridiculously easy for your competitors. And while there are ways to differentiate your GPT startup through branding and user interface, no company has as much to gain from using the technology as OpenAI itself.
Another concern about the emergence of text-generating systems is issues with output quality. Like many algorithms, text generators have the ability to absorb and amplify harmful prejudicesThey are also often astonishingly stupid. In tests of a medical chatbot built with GPT-3, the model responded to a “suicidal” patient by encourage them to commit suicideThese problems are not insurmountable, but they are well worth highlighting in a world where algorithms are already causing misarrests, unfair school grades and biased medical bills.
As OpenAI’s latest milestone suggests, GPT-3 will just keep talking, and we need to be ready for a world filled with robot-generated chatter.