What if I told you someone is taking a look at your photos to train their AI? Who? The old facebook.
FB has been training its AI by taking a look at your very own Instagram. While it may raise some brows regarding privacy issues, let’s take a further look at what it is all about.
Since the world is now progressing towards AI, new things are coming along the way so that business needs and customer requirements can be resolved.
FB is Looking at your Instagram pictures to train its image recognition AI regarding your activity and the trends of your preferences.
But again, these training will definitely bring about a situation that will require FB to clarify or improve their privacy policies. It is necessary to know about the choices and preferences of your customers for future sales, but taking a look at your photographs that are displaying your personal life so openly, to sell your data to the future advertisers definitely calls for a code of ethics.
How Does It Work?
FB has its “pre-training” that is mainly something like a research that is looking for relevant and synonymous hashtags. This is what they ended up terming as “ large-scaling hashtag traditional model”
Though it’s worth mentioning that FB is taking a look at only those accounts that are public on Instagram and not those that are ‘private’. But when a user posts a picture on Instagram are they aware that their data can be ever used up by an image recognition AI of a giant corporation for research purpose.
The Results
Well, we do not know what was FB’s intention behind searching on the hashtags, but we could understand that these researches are good only for FB in its sale through the future advertisers. The kind of results that were generated out of these researches was more objective than on a larger scale. Facebook cannot most definitely use the data generated out of these hashtags to get what you think about #tgif or #ogim. It can, however, take a look at the trends of your posts. For example, you could be a traveler or a makeup lover or someone who follows and reposts some authors. Your tastes and preference can definitely be seen through the relevant hashtags that you have been using or the similar hashtags that you are following.
The farthest that these researches went to somewhere around 3.5 billion images on Insta and that gathered up top about 17000 hashtags. Facebook exactly knew how to filter out those hashtags or the data that they could use for the research that was going to help them. Because looking at that figure, we don’t think any Megatech corp has that much of resource to study those staggering amount of data.
But what are your opinions about FB and its recent researches to train its image recognition AI? Do you see a rise in the need for a code of conduct? Let us know.