+5 votes
92 views
How to detect deepfakes and AI generated text?

in Guides by (551k points)
reopened | 92 views

1 Answer

+3 votes
Best answer

Politics, fertile ground
Simple tricks to detect deepfakes
More attention than fear
Find the source of the video
Capture and reverse image search
Analyze the context

We already talked about deepfakes at the time, and we wondered if we should be worried. Basically the answer was yes, due to its undoubted ability to deal damage in the real world. Well, this time we wanted to dwell on some of the methods that can be useful for detecting deepfakes, as well as other content generated by Artificial Intelligence..

The first thing to keep in mind is that we are talking about tools. As such, they can be used for purposes that do not pose any risk to the population. However, the danger of this being the case is latent. Precisely, this is because it is much easier to generate and spread false information.

Surely it has happened to you too: your parents, grandparents, siblings or friends tell you about something they have read on the Internet. Regardless of their knowledge of the web, they have taken it for granted. You search about it, and in just a few seconds you see that it is a hoax. However, that false notification has already affected those people, and who knows more. In some cases it may be a small error but, in others, it may have consequences..

If we add to that the fact that there are sites like Medium or Tumblr , where anyone with access to the network can post, then it is clear that we must be especially careful about what we read. And then there is the misrepresentation of traditional media, or the approach they provide according to their economic interests.

image

Politics, fertile ground

Without going into details about what a deepfake is, which we have already analyzed as we said, we are going to focus on some of the fields where they usually have more results. And there appears, in the undisputed first place, politics. After all, nowhere else is so much power and money wielded at the same time..

Just take a look on YouTube or other similar platforms, even on social networks, to see videos in which Barack Obama - and many other celebrities - supposedly say things that they never really said. If we pay attention, we will notice that there is Artificial Intelligence behind it. But the results are increasingly similar to reality. That makes many governments around the world fear about the reach of this technology.

On the other hand, there are those who have taken advantage of these advanced systems for not so spurious purposes. This is the case of the Indian politician Manoj Tiwari, who used the scope of this technology to translate himself into the different languages ​​spoken in that country, and thus convince a greater number of the population.

But, obviously, it is especially worrying that this kind of false notifications are used to discredit others, not so much to strengthen their own campaigns. Detecting deepfakes in such situations is even more of a priority. With these initial notions at hand, we are going to give you some tips that you should take into account.

image

Simple tricks to detect deepfakes

While a number of deepfake detection algorithms and tools are already in the works to help combat the problem, such as Reality Defender or Deepstar , there are a few suggestions you should consider:

Find the source of the video

One of the mechanisms that most often succeeds in the face of the possibility of a deepfake is to search for the source of the video. For example, if it is about Barack Obama, you can rest assured that a real message will always appear on his social networks. If it doesn't show up, you have a good reason to assume it's fake.

Similarly, it is convenient that you always distrust everything you see on social networks.

Capture and reverse image search

Some features of platforms that we use on a daily basis can also be useful against deepfakes. Without going any further, that is what happens with the Google image search engine. When you see news like this, the first thing to do is take a screenshot. Then do a reverse Google image search.

Analyze the context

Last but not least: always analyze the context. It will be difficult for a video in which the protagonist says something contrary to what he usually says to be true. For this reason, we advise you to always pay attention to recent previous statements, historical positions, and other data that can serve as a guide.

image

More attention than fear

In summary, we can say that deepfakes are a matter that we should be dealing with, rather than worrying about. As we adapt to it, and knowing its scope, we also develop systems that allow us to discover them faster and faster. Therefore, the key is to take advantage of these systems to quickly dismantle these hoaxes.

It would not be unreasonable, in fact, that at some point all online videos have a built-in detector, or something like that, that indicates if we are dealing with real content or one manipulated by Artificial Intelligence. Until that happens, it is best to distrust all sources, to avoid being deceived.

If something makes you suspicious, it is best to investigate thoroughly. If you are going to make decisions based on that information, it will be even more important. With this you will be safe from most of these deceptions.

And you, do you have any personal trick to detect deepfakes?



by (3.5m points)
edited

Related questions

+5 votes
1 answer
asked Apr 18, 2023 in Guides by backtothefuture (551k points) | 42 views
+4 votes
1 answer
asked Apr 19, 2023 in Guides by backtothefuture (551k points) | 36 views
+5 votes
1 answer
asked Apr 18, 2023 in Guides by backtothefuture (551k points) | 37 views
+5 votes
1 answer
asked Nov 27, 2020 in Help by backtothefuture (551k points) | 346 views
+3 votes
1 answer
asked Nov 18, 2019 in Mac by backtothefuture (551k points) | 330 views
Sponsored articles cost $40 per post. You can contact us via Feedback
10,632 questions
10,764 answers
510 comments
3 users