While working the other day, I came across an element within the spellcheck system of Grammarly which made me pause.
While writing about an island created in the Atlantic as a power plant, I used the term “manmade” to describe how the island was created.
This led to Grammarly informing me that the term “manmade” was gender-biased language and recommended that I change it to a more neutral term like “humanmade”.
Now, I understand exactly why Grammarly pulled me up on this. It is certainly a gender-biased term, but it got me thinking about the process which led to this error message being inserted.
Was it the AI itself which had decided that my gender-biased thinking needed to be changed? Or was it its programmers’?
If it was a decision made by the programmers, was it due to their own belief in the term needing to be corrected or as a precaution in case of backlash or outrage from users?
I then tried another few terms within Grammarly to see if they pass the PC test. “Mankind” is also gender-biased, but the saying “men from the boys” is not. “Manpower” is another no-no, but terms like “middleman”, “common man”, “rise of man”, and “mother tongue” are all perfectly acceptable.
Why Should Anyone Care About Gender-Biased Corrections?
Upon further research, I found a lot of outrage online about this decision. On most sites, I noticed that it was largely people against this change that were speaking out against this.
One article alone contained 52 comments all railing against Grammarly’s “authoritarian” approach to their writing style.
This raised the question of exactly why these people were becoming so annoyed at Grammarly’s gender-biased error suggestions.
While you could argue that gender bias should be a valid consideration in proper grammar, the primary argument against this feature focuses on a different aspect of the suggestion. That is, the program is telling the user what they should and should not say.
Naturally, people are resistant to being told what to do. If the directive already steps on their toes because of a perceived political association, as in the comment above, they’re going to go into defense mode.
In the past few years, the term “Political Correctness” has been thrown around with an infinite number of definitions. In most cases, however, it has been humans making the decisions on what is politically correct or not. Now, in our creation of Artificial Intelligence, we are giving these terms and notions to the programs we develop.
Although this seems like a minor issue when it comes to things like the use of the word “manmade”, the ethical programming of AI will become an enormous issue in years to come.
A clear example is autonomous cars. In the event of a crash, if both a pedestrian and driver have an equal chance of surviving, who does the car choose to save? What if it just decides to save itself?
As the Artificial Intelligence we create becomes more intelligent, we need to start discussing the issue of ethics a little more deeply. It may seem frivolous, but when AI is driving us to work, flying our planes, and even performing surgery on us, we need to know where the ethical and moral lines for these programs lie.
Political correctness is just the tip of the iceberg. However, it is an iceberg that is looming ahead, and one we need to focus on soon.