5G in Africa: What’s its potential, Selasi Ahorlumegah?
Naa Oyoe Quartey
With its fast speeds and revolutionary potential, 5G stands out as a noteworthy milestone in the field of […]
Is ChatGPT the world’s biggest thief?
Created by OpenAI's Dall-e AI tool
The New York Times is suing OpenAI, maker of ChatGPT and its biggest investor Microsoft, claiming that they illegally used content from the publisher’s site to train Artificial Intelligence systems.
The battle is the latest in a long running dispute between technology firms and media who have found their business models threatened by new developments.
The New York Times says OpenAI should pay it millions of dollars.
The arguments
The New York Times claims that ChatGPT uses material created by its journalists to respond to requests from users. In some cases, that material is not publicly available but behind a paywall and only distributed to subscribers. This amounts to a breach of copyright laws, the company says.
Artificial Intelligence companies have previously said they believe they have the right to train their models using material on the internet under an exemption in copyright law called “Fair use”. This is designed to prevent vital information from being owned or monopolised by a single company.
The facts
AI models derive their power from the information used to train them. The latest iteration of so-called Generative AI is based on Large Language Models which are able to “understand” language thanks to having studied huge amounts of writing. This can come from sources such as news reports, books, or even recipes.
Novelists including John Grisham and record labels like Universal Music have already begun court cases against AI companies for using their material without permission.
In response, some companies that create AI products including Google have promised to protect their users from copyright lawsuits resulting from their products.
The battle between Big Tech with the media industry is a long-running one. There have been disputes about how social media and search engines use news reports to make money without paying its creators for many years. The loss of revenue suffered by media companies as a result of Facebook, Google and others coming to dominate advertising markets has led to many thousands of job losses.
Some governments have been persuaded to introduce laws forcing tech companies to pay for news media. However, the companies claim that by providing a service to users they have helped distribute news stories more widely.
In the domain of AI, some companies have blocked AI models from accessing their material, others such as German giant Axel Springer have done deals to get paid for providing content to the tech industry. Media are also using AI models – sometimes developed in house, sometimes based on ChatGPT or other products – to create content, creating concerns of a giant loop where AI is creating the content that is being used to train new AI.
