Trends show humanity could reach singularity in just 6 years

• By one unique measure, we could be approaching the technological singularity by the end of this decade, if not sooner.

• The translation company developed the Time to Edit (TTE) metric to calculate the time it takes professional human editors to revise AI-generated translations compared to human translations. This can help quantify the speed of reaching the singularity.

• Artificial intelligence that can translate speech just like humans can change society.

In the world of artificial intelligence, the idea of ​​“singularity” is of great importance. This slippery concept describes the moment when AI moves beyond human control and rapidly transforms society. The trick is AI’s singularity (and why it’s borrowing terminology from black hole physics) is that it is extremely difficult to predict where it will start, and almost impossible to know what lies beyond this technological “event horizon”.

However, some AI researchers are looking for signs of reaching the singularity, measured by AI progressing towards human-like skills and abilities.

One such metric, identified by Rome-based translation company Translated, is the ability of AI to translate speech with human accuracy. Language is one of AI’s most challenging problems, but a computer that could fill this gap could theoretically show signs of artificial general intelligence (AGI).

“This is because language is the most natural thing for humans,” Translated CEO Marco Trombetti. said at the conference in Orlando, Florida in December 2022. “However, the data collected by Translated clearly shows that machines are not that far away from closing the gap.”

The company tracked the performance of its AI from 2014 to 2022 using a metric called Time to Edit, or TTE, which calculates the time it takes professional human editors to revise AI-generated translations versus human translations. Over this 8-year period and analysis of more than 2 billion After editing, Translated’s AI showed a slow but undeniable improvement, gradually closing the gap with human-level translation quality.

According to Translated, on average it takes a human translator about one second to edit each word of another human translator. In 2015, it took professional editors approximately 3.5 seconds per word to check a sentence translated using machine translation (MT), but today that number is only 2 seconds. If this trend continues, by the end of the decade (or even sooner) Translated’s AI will be as good as human translation.

“The changes are so small that you don’t notice them every single day, but when you see the progress… over 10 years, it’s impressive,” Trombetti said. on the podcast. “This is the first time that anyone in the field of artificial intelligence has predicted the speed at which the singularity will be reached.”

While this is a new approach to quantifying how close humanity is to the singularity, this definition of singularity faces similar problems. broader definition of AGI. And while improving human speech is certainly the cutting edge of AI research, impressive skills don’t necessarily make a machine intelligent (not to mention how many researchers I don’t even agree about what “intelligence” is).

Whether these ultra-precise translators are the harbingers of our technological doom or not, this does not detract from Translated’s achievements in the field of artificial intelligence. AI capable of translating speech like humans could very well transform society, even if a true “technological singularity” remains elusive.

You may also like