Anicet Mbida 06:54, September 30, 2022

Anicet Mbida delivers to us every morning what is best in terms of innovation.

This Friday, he is interested in the ease with which one can now create fake images and fake videos thanks to advances in artificial intelligence.

More than ever, we will have to be wary of what we see on the internet.

It has never been so easy to fabricate fake images and even fake videos.

All this, once again, thanks to advances in artificial intelligence.

Yesterday there were two big announcements that completely changed the game.

The first is the provision to the public of a tool called DALL-E.

It is a software capable of generating any image simply from keywords.

You tell him, for example, "a teddy bear that juggles on the wings of an airplane" and he will make you several perfectly similar proposals.

You can even start from an existing photo or your own photo and ask “me, disguised as an cosmonaut in a bathtub full of kiwis”.

It's childishly simple.

It's free.

And the results are impressive.

It's still worrying, does that mean that we can make fake news more easily?

This is the problem.

Note, however, that this is a research tool.

It is developed by OpenAI, an NGO that wants to ensure that advances in artificial intelligence will first benefit humanity.

Some terms are therefore filtered (Ukraine, for example, to avoid fake news).

You can't generate obscene images either… But you're right, there are risks.

Typing three keywords is easier than mastering Photoshop.

Especially since we have already moved on to the next step.

A little later in the day yesterday, it was the Meta researchers, this time, who presented a similar tool… but capable of generating videos.

There again, it is almost magical: “a dog with a flying superhero cape” and we get a video of a labrador playing Superman.

The tool is called Make-A-Video.

I put examples on the hashtag #Europe1.

Go see, it's stunning!

Also, is it available?

Anyone can try it?

Nope !

For the moment, it is reserved for researchers.

They are aware of the potential abuses, so they are working on safeguards before opening it to the public.

And all of this is just the beginning.

With artificial intelligence, we have clearly just passed a milestone… For better and perhaps for worse.