Artificial intelligence amazes: – Pioneering

Imagine that Erling Braut Haaland is reading you an audiobook.

Or that the late Kurt Cobain of Nirvana writes and sings new lyrics.

This may be possible soon.

Google launched a new one Artificial Intelligence (AI) who runs what he belonged to a short time ago in a science fiction movie.

AudioLM, as the AI ​​calls it, can hear a short beep – and then create its own continuation.

Give him three seconds of an audiobook – the AI ​​continues to read in the same voice. What you say makes up for itself.

Give it three seconds of a piece of music – the AI ​​continues to compose with the same instrument.

According to the researchers behind the sounds and voices, AI makes it so good that people struggle to hear that it’s not real.

Hear for yourself in the video above. Are you fooled?

– Nice

Particle physicist and artificial intelligence expert Inga Stromki believes innovation is groundbreaking.

– I think it’s very cool. I didn’t doubt it would happen, but I didn’t expect it to happen so quickly.

Many devices are already capable of generating logical text. Robotic voices have also been with us for a long time. Maybe you know Siri or Alexa?

AudioLM can generate the sentences themselves. But he also runs something that would normally have been impossible for machines: natural intonation.

The AI ​​researcher is amazed.

– Creating a system that can constantly create content, and then produce the tone and sound that suits it, testifies to the interest of the work, says Inga Stromki.

You can hear all the examples of what AudioLM can do here.

Tone occurs naturally in humans. We give the meaning of our language by using a certain tone of voice.

Instrumental voices, for example, sound artificial precisely because the tone of voice does not adapt to what is being said.

Stromki explains that AudioLM seems to understand roughly what it’s saying per se.

What sound he makes depends on how he interprets the words, she says.

Audio books and music

Google confirms that AudioLM is still in the research phase. They don’t want to fully release the technology yet.

Strümke believes that innovation can lead to major developments in many areas in the future.

– It is conceivable that dubbing movies and reading audiobooks will not be a profession for people, she says.

With enough data, it can recreate what it should be. The voice of Erling Braut Haaland, for example.

Strömke believes that AI can also shape the future of music.

We can imagine that we can get artificially intelligent music artists, who will have a much greater ability to compose music than human artists, says the AI ​​expert.

Ethical challenges

Machines that speak or play music, like humans, can provide tremendous opportunities for creativity and development.

But should we also care about developing this kind of intelligent machine?

No, says the expert. At least not in this case.

AudioLM cannot make decisions for us. The goal is to imitate something very easy for humans to do, which is to adapt the tone of voice to what we say, says Stromki.

Christoph Trattner, Professor and Center Director at Media Futures at the University of Bergen, explains that new technology always brings with it new ethical issues.

— It’s always about how technology is used, he says.

Trattner explains that even if the innovation was intended for good, someone would want to use it for something evil.

Whether the use of AudioLM can lead to fraud, plagiarism or fake news remains uncertain.

– It’s very hard to say. It opens up a new ethical field. to new challenges, says Trattner.

Then it will be up to us to develop ethics in the region and find solutions to the challenges, he adds.

man and machine

Inga Stromki stresses that developments in the field of artificial intelligence are moving very quickly.

One of the challenges that humans face is when machines start solving problems for us. Without saying how and why.

It is happening more and more often in our time.

So long as we don’t understand what the machines have understood, we see you as a great intellectual debt, she says.

It’s great that computers outperform Magnus Carlsen at chess, but it’s great that we learn something new about chess from machines.