Breakthrough implant creates speech from brain signals alone

26 Apr 2019

The brain-computer interface could give speech to those who have lost it. Image: UC San Francisco

This week in future tech, researchers have created a device that can create speech just by thinking it, while China wants to ban ‘deepfakes’.

It is early days, but results from a recent study conducted by neuroscientists at UC San Francisco (UCSF) have shown amazing promise for a brain-computer interface that can generate natural-sounding synthetic speech by using brain activity to control a virtual vocal tract.

This anatomically detailed computer simulation includes the lips, jaw, tongue and larynx, and could one day restore the voices of people who have lost the ability to speak due to paralysis or other forms of neurological damage.

“For the first time, this study demonstrates that we can generate entire spoken sentences based on an individual’s brain activity,” said Edward Chang, a professor of neurological surgery and member of the UCSF Weill Institute for Neuroscience.

“This is an exhilarating proof of principle that with technology that is already within reach, we should be able to build a device that is clinically viable in patients with speech loss.”

However, it is by no means a perfect tool of communication just yet as, while the virtual vocal tract can reproduce the sound of someone’s voice, it will often include garbled words.

Still, a major accomplishment has been achieved, and the researchers hope to improve it over the coming years. Their findings have been published to Nature.

‘Deepfakes’ may be made illegal in China

China Daily has reported that the notorious use of ‘deepfakes’ – where artificial intelligence (AI) collects video and audio recordings of people, and then uses them to create false representations – will likely be made illegal in the one-party state.

The country’s top legislators recently raised concerns about the technology, with a draft law being deliberated by the Standing Committee of the National People’s Congress to make it a violation of individual portrait rights.

The draft law states that no person or organisation can distort another person’s picture or mimic their voice using AI, or any technology for that matter. One senior legislator said that deepfakes not only affect individuals, but the practice “also harms national security and the public interest”.

The country’s biggest deepfake incident so far was seen in February when the face of actor Athena Chu was replaced with that of another big celebrity in China, Yang Mi, as part of a TV show. The incident quickly went viral in China, drawing concern from the nation’s government.

Lasers may open the door to ultrafast Wi-Fi

Researchers from the Harvard John A Paulson School of Engineering and Applied Sciences have demonstrated a semiconductor laser that can emit microwaves wirelessly, modulate them and receive external radio frequency signals.

The senior author of the study, Federico Capasso, said that this research “opens the door to new types of hybrid electronic-photonic devices and is the first step toward ultra-high-speed Wi-Fi”.

The research builds on previous work that found an infrared frequency comb in a quantum cascade laser could be used to generate terahertz frequencies that can move data hundreds of times faster than today’s wireless platforms.

A later discovery showed quantum cascade laser frequency combs could also transmit and receive encoded information. Now, the researchers have found a way to extract and transmit wireless signals from these combs.

First author of the paper, Marco Piccardo, said: “This all-in-one, integrated device holds great promise for wireless communication.

“While the dream of terahertz wireless communication is still a ways away, this research provides a clear roadmap showing how to get there.”

Ford partners with Amazon to develop cloud services for autotech

Car manufacturer Ford – through its connected vehicles subsidiary, Autonomic – has teamed up with Amazon Web Services (AWS) in one of the latest major collaborations between US big tech and the auto industry.

The multiyear global agreement will reportedly “expand the availability of cloud connectivity services and connected car application development services for the transportation industry”. As part of the deal, Autonomic’s cloud service will be powered by AWS and will create partner and business opportunities for automakers, public transport operators, large-scale fleet operators and software developers.

“One of the many goals we have in common is to build a cohesive framework for vehicle makers and developers, allowing them to focus on creating some of the best experiences for customers who use our connectivity systems,” said Gavin Sherry, co-founder and CEO of Autonomic.

Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.

Colm Gorey was a senior journalist with Silicon Republic