Hillary Clinton and Donald Trump. (Photos: Getty Images)
Although the technology that is used for artificial intelligence and machine learning has been around for decades, only in the last few years is starting to get mainstream acceptance. For example, the iPhone has, since 2011, equipped with the intelligent personal assistant Siri, used for the understanding of speech. Android devices are Google Assistant, and a speech-to-text tool with the name “Voice typing” that can be used in combination with Google Docs.
Trint, which provides an AI-based software to transcribe recorded audio can be copies of 98 percent correct in a few minutes and at low cost, according to the company.
But the transcribing of the ramblings of a politician is another matter.
FORMER GOOGLE AND UBER ENGINEER IS DEVELOPING AN “AI GOD”
“Trint is not dictation software, but rather speech-to-text technology that works from a large corpus of training data,” Jeff Kofman CEO and Founder of Trint told Fox News. And the human part of the equation. “Trint is a man-machine hybrid model: we push the boundaries of what A. I. technology is able to with the help of man to the last pass if necessary,” he added.
Recently, Trint set a challenge to see which prominent politicians are the easiest for AI bots to understand. The challenge covered 14 “influential” politicians, among them President Trump, former Secretary of state Hillary Clinton, the US Ambassador to the United Nations Nikki Haley, the former President of the usa, Barack Obama Sr. Ted Cruz, R.-TX., Sen. John McCain, R.-AZ. and fictional politician Selena Meyer, played by actress Julia Louis-Dreyfus on HBO’s “Veep.”
The company took audio recordings of their speeches and by means of the transcription software. The results were analysed using the Word-Error-Rate (WER) as standard, often used to measure the accuracy of speech recognition. A transcript with a WER of 4.8, for example, is 4.8 percent incorrect, or 95.2 percent correct.
Who scored the highest?
Nikki Haley came in the first, according to Trint, with an accuracy rating of 99.48 percent.
Former Secretary Clinton was second, with a score of 99.38 percent, while Louis-Dreyfus “Veep” character, Meyer, came in fourth, scording 98.74 percent.
“It is worth noting that Julia Louis-Dreyfus is an experienced actress, so it is not surprising that her speech is clear and clean,” Trint added.
FAKE NEWS IS HERE TO STAY, SAYS STUDY
Who scored the lowest?
Trint said the lowest scoring politicians had a number of common speech patterns.
Both House Minority Leader Nancy Pelosi and Trump have the tendency to swallow parts of a word when they speak, Kofman told Fox News.
In some cases, heavy accents pain of the results. Thanks to Cruz, the Texas accent, he 13th place.
President Trump 11th at 97.89 percent. “Thanks to the slow, measured way of speaking, President Trump has a much lower overall word count than is usual for a politician; 380 (for this example, taking the average over all speakers is 486,” Kofman said.
“In addition, he has a tendency to swallow some of the prefixes and suffixes of words; for example, the word ‘incredible’ in the expression ‘That is something. And I am honored by your incredible support and are grateful for your commitment to our common cause’ was shown as ‘credible’, the Kofman went on to say.
Pelosi came in last among the 15 politicians, with a 96.01 percent accuracy, according to Trint findings.
Kofman added that the accuracy can also be affected simply by the recording quality – that has nothing to do with a person’s speech patterns. For example, some errors occurred because the Trint system missed words that were spoken away from the microphone.
“Although Trump and Pelosi was lower than others, they still come very good in these samples for the Trint Index,” Kofman said. “The human transcription’t do much better than that and take much more time and cost a multiple more.”