Terms of Endearment

No, this article is not about the 1983 Shirley MacLaine-Jack Nicholson box-office hit; it is about OpenAI, the latest poster child for exploiting your data.

If you, like millions of others across the globe, use Google, Microsoft or some other company website on a regular basis, I am sure you don’t remember or don’t care what exactly you agreed to when you signed on to use the services provided by these companies. Should you worry?

Consider the following ‘Terms of Use’ clauses from the OpenAI website.

Read | Going places with an iPhone, Starbucks coffee cup in hand

“…We and our affiliates own all rights, title, and interest in and to the Services...We appreciate feedback, comments, ideas, proposals and suggestions for improvements. If you provide any of these things, we may use it without restriction or compensation to you…You may not…reverse assemble, reverse compile, decompile, translate or otherwise attempt to discover the source code or underlying components of models, algorithms, and systems of the Services..; use output from the Services to develop models that compete with OpenAI; …use any automated or programmatic method to extract data or output from the Services, including scraping, web-harvesting, or web data-extraction; …You may use Services only in geographies currently supported by OpenAI”.

Since you are anxious to try out OpenAI’s ChatGPT, you are predisposed to clicking ‘yes’ to the terms of service without bothering to read through the documents that explicitly spell out the terms of use and privacy policies of OpenAI. A bad move on your part, I must say.

The hidden implications of the key words and phrases in the clauses just quoted are mindboggling. For starters, you can use an OpenAI product but do not question, by whatever mechanism, the results that are generated. If you do, you are violating the intellectual property rights of OpenAI, never mind that you unwittingly signed away your own IP by interacting with the product and are subject to sanctions imposed by the US-based company but, more importantly, the US government. It is the latter that is more problematic since the US ensures IP protection rights are enshrined in all trade treaties, violations of which can trigger global sanctions at the individual and country levels. It is no coincidence that China, Cuba and North Korea are on the short list of countries where OpenAI is not supported.

The global wrangling on data colonisation has focused on where data should be stored, but this is a mere distraction. While ‘right to be forgotten’ legislation can be used to specify the length of time a website is allowed to store user data, it is a lot trickier to come up with suitable laws for dealing with how user data is to be used.

Please bear in mind that private companies set their own rules. Here is an egregious terms of service example: “You hereby grant to the Washington Post a royalty-free, irrevocable, perpetual, worldwide, non-exclusive, and fully sublicensable license to use, reproduce, modify, adapt, publish, translate, create derivative works from, incorporate into other works, distribute, perform, display, and otherwise exploit such content that you provide, in whole or in part in any form, media or technology now known or later developed”. [italics are mine]

It is okay for OpenAI to harvest web data but you, on the other hand, are constrained to treating OpenAI as a black box. When anonymised medical transcriptions of your and numerous other patient-doctor interactions are used by OpenAI’s learning algorithms to pinpoint the etiology of a disease and come up with a potential cure, the ‘cure’ belongs to the company, not the world at large.

The call by the five Big Tech companies for a six-month moratorium on creating advanced AI software has nothing to do with a sudden realisation of their newfound love for humanity but acute concern for their own survival in the marketplace of innovative products. Search engines such as Google and Bing generate their output based on information that already exists in various forms on the web. Incorporating OpenAI’s algorithms into these search engines will lead to extracting new knowledge from existing data and this is going to prove problematic for the IT behemoths. According to their service terms, OpenAI can claim ownership of the newly created knowledge/information.

A more dangerous possibility is one in which OpenAI can lay claim to any new knowledge created anywhere in the world by anyone with internet access and demand payment.

The company with the Midas touch. Goodbye creativity, hello greed.

Honi soit qui mal y pense.

(Roger Marshall, a computer scientist, a newly minted Luddite and a cynic.)