Google Has Developed An AI That Can Lip-Read Better Than Humans!


As you know, Google Translate has recently started translating complex texts surprisingly well. Translators all over the world were almost terrified by the success of an AI, but it seems Google won’t stop developing AIs that can take over many other jobs. This time they’ve created a lip-reading system that leaves professionals in the dust.


A project by Google’s DeepMind and the University of Oxford applied deep learning to a huge data set of BBC programmes to create a lip-reading system.

The AI was trained using 5000 hours of videos from six different TV programs, including Newsnight, BBC Breakfast and Question Time.

By looking at the speaker's lips, the AI accurately deciphered complex sentences like “We know there will be hundreds of journalists here as well” and “According to the latest figures from the Office of National Statistics.”

The system has even beaten a professional lip-reader. The professional annotated just 12.4 percent of words without any error. But the AI annotated 46.8 percent of all words in the data set without any error.

Many of its mistakes were just small slips.

What’s more is that the AI could even read videos with sync problems.

It was taught the correct links between sounds and mouth shapes. Using this information, the system figured out how to match up, and realigned them successfully. It automatically processed all 5000 hours of the video and audio, which is almost an impossible task for us humans.

The question is now how to use DeepMind’s incredible abilities.

“We believe that machine lip readers have enormous practical potential, with applications in improved hearing aids, silent dictation in public spaces (Siri will never have to hear your voice again) and speech recognition in noisy environments,” says computer scientist Assael.

There are, of course, privacy concerns.

Imagine the world under constant surveillance by the CCTV cameras and the smartphone in your pocket, but there will also be an AI out there that can read your lips even without actually hearing you.

Nevertheless, it’s still astonishing to see such groundbreaking developments in technology, so let’s just hope they won’t take over all the jobs in the world someday.

How do you feel?
Tears of Joy
Relieved Face
Clapping Hands
Thumbs Down
Send Feedback