2 Matching Annotations
  1. Last 7 days
  2. www.pewresearch.org www.pewresearch.org
    The challenges of using machine learning to identify gender in images
    2
    1. ieh12 06 Nov 2026
      For example, a model trained on images in which every man is wearing glasses might be convinced that wearing glasses is a strong signal that someone is a man. While this example might seem obvious, it can be difficult to know in advance if training data consists of poor examples for the task, or if it is unrepresentative.

      This example made me think about how easily technology can learn something incorrect if the information used to train it is limited. Wearing glasses has nothing to do with being a man or a woman, but the system might make that assumption. I believe this shows why we should not always trust the answers provided by artificial intelligence. It also makes me wonder how many other incorrect assumptions these systems might make without us noticing. In my opinion, the people who develop these technologies have a responsibility to use diverse information and carefully check the results.

    2. ieh12 06 Nov 2026
      What may be less obvious is that they can be significantly less reliable for some groups than others – and that these differences may not necessarily be driven by intuitive or obvious factors. In general, it is important that these models be trained on data that captures the diversity of the situations they will encounter in real-world contexts, as much as that is possible. If the model will be called upon to operate on multiple ages, races and other qualities, for example, it is important that the model be trained on a similarly diverse training

      I think this is an important issue because technology should work fairly for everyone, regardless of their race, age, or gender. It concerns me that some people may receive less accurate results simply because the technology was not trained with enough diversity. I believe this could create unfair situations, especially when these systems are used to make important decisions about people. It also makes me wonder who is responsible when AI makes mistakes that affect certain groups more than others. In my opinion, companies should test their systems with people from different backgrounds before making them available to the public.

    Visit annotations in context

    Annotators

    URL

    pewresearch.org/internet/2019/09/05/the-challenges-of-using-machine-learning-to-identify-gender-in-images/