English grammar can be tricky so that even the smartest people who are native make mistakes. This is why using a grammar checking software might seem like a good idea. This idea is supported by aggressive advertising by leading services in this field.
But real life is a bit more complicated, and users often face some unexpected problems. Today we will talk about these issues, and how to determine a grammar checking tool that will deliver better experience and results.
Robots want to outsmart humans
One of the most significant problems of writing correction tools is that they pretend to be smarter than their users. However, the current level of technological development does not allow that.
For example, take Grammarly, the leading grammar checking tool on the market. Its algorithm hates passive voice and trying to eliminate it everywhere possible in 100% cases. The system considers passive voice a bad writing construction as opposed to active voice.
Here is how this looks in practice. Let’s suppose that you have a sentence: “The car was parked by David.” The grammar checker will highlight an error — sure thing. There is a passive voice! A sophisticated algorithm will come up with a rewritten sentence: “David parked the car.” Sounds good so far, but what if your sentence is “The car was parked by the sidewalk”? Here we mean that the car is physically parked close to some object (a sidewalk in our case). And what an AI-based algorithm worth of $29.95/mo offers here?
Here you are:
This is what happens when the computer tries to outsmart human. Online helpers should be less aggressive, not try to be too strict about generated recommendations, no matter what they relate to from passive voice to starting sentences from And or But.
For example, a built-in MS Word grammar checker used to be like Grammarly a couple of years ago, but now it works in a softer manner and does not burn out passive voice, for example.
Devs overlook privacy
Recently Sebastian McKenzie, the Facebook engineer, has raised an interesting topic: the degree of privacy grammar checkers allow. He called Grammarly a keylogger because of its maniacal desire to log any information the user types on the web (excluding restricted fields like passwords forms).
Here's a summary of why Grammarly are predatory and why you should avoid them. I am not a lawyer™. This is my layman reading, but they are pretty explicit about a lot of these practices. Posting because I've seen a lot of people oblivious to what they're signing away.
Funny enough that a Facebook employee raises privacy concerns, but he mentioned the real facts. You can’t call grammar checker like Grammarly a keylogger as the users grant the permissions and installs browser extension by his own, but the question is why this tool gives no control over privacy to its users?
If you do some research, you will find other tools with some privacy-related features. For example, Textly.AI allows using its Chrome and Firefox extensions in an incognito mode without registration and features a pastebin mode which works like secret chat in messaging app in its web app. This is a much better approach for users. But still, you need to search for such tools by yourself.
Software does not help in content creation
Another serious problem of many grammar checkers is that these tools check what was already written. But a vast amount of people even do not get that far: ask Google about writer’s block, and you’ll get ~29 million of search results. This is a real problem modern tools do not help to solve.
Little is done in this direction now. One of the solutions: Textly.AI has released its content templates library so that users could build their content on top of it:
Another example: Ginger allows a sentence rephraser that is capable of creating more sophisticated sentences from the simple ones. Still it can’t fix the core problem of the writer’s block.
Modern grammar checkers and writing enhancement tools can be useful. However, this class of software has its limitations and problems. Thus, to choose a tool that will help you solve your task, there are several questions to answer:
- Do developers work on privacy, is it one of their goals?
- Does the service helps with content creation, or fix what you’ve already written?
- Is algorithm soft enough to do not force you into using irrelevant recommendations?
That’s it, thanks for reading and post your thoughts on the topic below!