Google Docs users are sharply critical of artificial intelligence algorithms used for automated typing

As language models are built into some apps, new assisted typing features will be added to Google Docs to help users type faster and better.

And this is that in addition to grammatical and spelling corrections, We will see suggestions to improve our writing style as we write.

Assistive input is introduced primarily for business users, and this feature is enabled by default. The problem is that not everyone is in favor of being guided by an algorithm, and some people find his ability for nonsensical, inclusive languages ​​annoying.

Some examples we found are that words like “cops” can cause the model to suggest changing to something more neutral like “cops”. Another case is the replacement of “landlord” with “owner” or “owner”, or the automatic elimination of profanity.

Although this new system has a positive effect on the improvement of writing, To what extent does it really fulfill this function and harm rather than improve?

There is a major problem that auxiliary writing uses language comprehension models that based on millions of common phrases and sentences to automatically learn how people communicate. However, this also means that they may reflect some human cognitive distortions.

By this we mean that while experts strive to create a perfect intelligent system, lFairness in machine learning models is difficult to measure and improve.

Trying to do justice is difficult, not only because people don’t agree on what the term means, but also because solutions for one application may not be suitable for another.

“We have some problems. We have to decide ethics. We must make sure that the mistakes of the past are not repeated. We have to understand the life science of AI. Otherwise, we will create a monster. very optimistic and I think that if we pay attention, we can solve all these issuessays IBM CEO Arvind Krishna.