News
Scientists use AI neural network to translate speech from brain activity
Three recently published studies focused on using artificial intelligence (AI) neural networks to generate audio output from brain signals have shown promising results, namely by producing identifiable sounds up to 80% of the time. Participants in the studies first had their brain signals measured while they were either reading aloud or listening to specific words. All the data was then given to a neural network to “learn” how to interpret brain signals after which the final sounds were reconstructed for listeners to identify. These results represent hopeful prospects for the field of brain-computer interfaces (BCIs), where thought-based communication is quickly moving from the realm of science fiction to reality.
The idea of connecting human brains to computers is far from new. In fact, several relevant milestones have been made in recent years including enabling paralyzed individuals to operate tablet computers with their brain waves. Elon Musk has also famously brought attention to the field with Neuralink, his BCI company that essentially hopes to merge human consciousness with the power of the Internet. As brain-computer interface technology expands and develops new ways to foster communication between brains and machines, studies like these, originally highlighted by Science Magazine, will continue demonstrating the steady march of progress.
In the first study conducted by researchers from Columbia University and Hofstra Northwell School of Medicine, both in New York, five epileptic participants had the brain signals from their auditory cortexes recorded as they listened to stories and numbers being read to them. The signal data was provided to a neural network for analysis which then reconstructed audio files that were accurately identified by participating listeners 75% of the time.
In the second study conducted by a team from the University of Bremen (Germany), Maastricht University (Netherlands), Northwestern University (Illinois), and Virginia Commonwealth University (Virginia), brain signal data was gathered from six patients’ speech planning and motor areas while undergoing tumor surgeries. Each patient read specific words aloud to target the data collected. After the brain data and audio data were given to their neural network for training, the program was given brain signals not included in the training set to recreate audio, the result producing words that were recognizable 40% of the time.
Finally, in a third study by a team at the University of California, San Francisco, three participants with epilepsy read text aloud while brain activity was captured from the speech and motor areas of their brains. The audio generated from their neural network’s analysis of the signal readings was presented to a group of 166 people who were asked to identify the sentences from a multiple choice test – some sentences were identified with 80% accuracy.
While the research presented in these studies shows serious progress towards connecting human brains to computers, there are still a few significant hurdles. For one, the way neuron signal patterns in the brain translate into sounds varies from person to person, so neural networks must be trained on each individual person. The best results require the best data possible, i.e., the most precise neuron signals possible, meaning this is something that can only be obtained by placing electrodes in the brain itself. The opportunities to collect data at this invasive level for research are limited, relying on voluntary participation and approval of experiments.
All three of the studies highlighted demonstrated an ability to reconstruct speech based on neural data in some significant capacity; however, also in all cases, the study participants were able to create audible sounds to use with the computer training set. In the case of patients unable to speak, the level of difficultly in interpreting the brain’s speech signals from other signals will be the biggest challenge. Also, the differences between brain signals during actual speech vs. thinking about speech will complicate matters further.
For years, there have been images and videos across social media platforms that have reminded me of when I was a 15-year-old kid teased by “Xbox 720” videos on YouTube. These videos are of the supposed “Tesla Phone” that Elon Musk was secretly developing in between leading Tesla with its electric cars and SpaceX with its reusable rockets.
Would you buy a Tesla phone ? pic.twitter.com/aaTwvvIJit
— Tesla Owners Silicon Valley (@teslaownersSV) October 6, 2023
Although Musk has put those rumors to bed several times, it was never completely out of the realm that he could get involved in cell phones in some capacity. Think outside the box and more macro-level, though. Instead of reinventing the computer, Musk reinvented connectivity by developing Starlink with SpaceX.
It could be something similar, TD Cowen analyst Gregory Williams said in a note last week, where he hinted SpaceX could be gathering some steam to acquire T-Mobile.
Williams said it would be the “clear choice” for SpaceX if it decided to go through with a network acquisition. He also suggested AT&T.
The move would be possible through selling more of its own stock, which would help SpaceX raise the money to purchase T-Mobile, which would cost roughly $300 billion. It could be one of the moves SpaceX makes post-IPO in terms of an acquisition: it already acquired Cursor AI for $60 billion.
Other analysts, like Dan Ives of Wedbush, believe SpaceX and Tesla will eventually merge into one anyway, and that conglomeration could come as soon as this year, some have said.
The implications of SpaceX purchasing T-Mobile are massive. A combined entity would create a truly ubiquitous network: T-Mobile’s terrestrial 5G towers and Starlink’s growing constellation of Direct-to-Cell satellites. This would essentially eliminate dead zones across the U.S. and potentially globally.
SpaceX would instantly become a full-scale facilities-based carrier with satellite differentiation; a huge advantage. This would pressure AT&T and Verizon heavily.
There are also concerns like a potential reduction in long-term competition, and of course, a deal of that size would face intense scrutiny from government agencies.
The strategic fit is compelling due to the existing Starlink–T-Mobile partnership and complementary technologies (space + terrestrial). It could create a dominant integrated communications player. However, the regulatory, financial, and execution hurdles are enormous — this remains highly speculative with no indication SpaceX is actively pursuing it right now.
Tesla revealed a major new Cybercab detail in a guide it released for First Responders, showing new territory in its beliefs and intentions for the ride-hailing-focused vehicle that entered production in April.
The First Responders Guide is released to give fire departments, paramedics, and other emergency personnel the proper guidance on what to do in the event of an accident, entrapment, or other situation that would require immediate attention.
On one of the pages of the First Responders Guide, Tesla revealed a stark detail about the Cybercab, which could help personnel enter the vehicle more easily in case of an emergency.
Tesla Cybercab has one important piece that AI4 cars might need for FSD
It shows Tesla has no intention of releasing any Cybercab units that were initially proposed for ride-hailing services for the general public with any manual controls, meaning a steering wheel or pedals:
“A Cybercab equipped with steering wheel, brake pedal, and an acceleration pedal is typically an engineering or test vehicle, and operates at SAE Level 2 autonomy. Cybercab is not typically equipped with a steering wheel or acceleration and brake pedals.”
New official Cybercab documentation from Tesla:
“A Cybercab equipped with steering wheel, brake pedal, and an acceleration pedal is typically an engineering or test vehicle, and operates at SAE Level 2 autonomy. Cybercab is not typically equipped with a steering wheel or… https://t.co/P6ut1mZyzr pic.twitter.com/yq6skl9s2J
— Sawyer Merritt (@SawyerMerritt) June 27, 2026
This is a major development for those who continue to believe Tesla planned to release the Cybercab with any sort of manual controls so that passengers could take over if needed. However, when Tesla started manufacturing production versions of the Cybercab in Giga Texas earlier this year, they were spotted without a steering wheel or pedals.
It essentially confirms the company has no intentions of bringing manual controls to the car’s production versions. Some have argued that the likelihood of Tesla having something
There still are some Cybercab units out there with a steering wheel and pedals, and as Tesla said, these cars are engineering or test vehicles, which have Safety Monitors on board to help the car out of a precarious situation or emergency.
Tesla released the Full Self-Driving v14 ‘Lite’ suite to owners of Hardware 3 or AI3 vehicles today, adding several new features to the vehicles that were once believed to be capable of unsupervised self-driving.
Now, Tesla has released this modified suite to older Tesla vehicles, adding plenty of new features and capabilities.
Here are the full release notes for the suite:
- Distilled the intelligence from HW4 V14 into HW3. This allows HW3 to directly learn how to handle scenarios using HW4 V14 as a guide. This process unlocks the improvements that have been made to HW4 including Reinforcement Learning (RL) and offline models for HW3.
- Improved both proactive and reactive responsiveness across a wide variety of categories including navigation handling, merges and forks, pedestrian interactions, traffic lights, and vehicle cut-in scenarios.
- Improved general comfort in nominal scenarios through fewer false slowdowns, smoother steering and more consistent lane centering.
- Introduced parking, unparking, and reversing capabilities.
- Added Arrival Options for you to select where FSD should park: in a Parking Lot, on the Street, in a Driveway, or at the Curbside.
- Speed Profiles are now available at all times, to further customize driving style preference.
These improvements, according to Tesla’s Head of AI, Ashok Elluswamy, help distill the driving behavior from AI4’s v14 series into both the camera and compute configurations of AI3.
Tesla Full Self-Driving v14 ‘Lite’ for older cars finally gets released
He added:
“It includes destination options and speed profiles on city roads, but more importantly significantly improved safety. We hope you’ll enjoy it, once the build ships wide.”
FSD v14 Lite is now rolling out to AI3 early-access customers. Based on the feedback, will rollout to more customers over the next few weeks.
This build distills the driving behavior from AI4’s v14 series into both the camera and compute config of AI3. It includes destination…
— Ashok Elluswamy (@aelluswamy) June 29, 2026
Tesla will continue to roll out the v14 Lite suite more widely in the coming weeks, the company said.
Tesla Phone? Not quite, but close: analyst
Tesla reveals huge Cybercab detail in new guide for First Responders
