Connect with us

Tech

Scientists succeed in ‘mind-reading’ using ChatGPT

Published

on

With the help of artificial intelligence (AI) powered ChatGPT, neuroscientists believe they have found a way to translate the activity of the brain into words, a major discovery that can help patients with conditions like “locked-in” syndrome, stroke, etc that render them unable to communicate.

The scientists from the University of Texas in Austin used the groundbreaking OpenAI’s human-like chatbot showing its applications in the healthcare sector as AI is on the way to modernisation and advancement, eventually touching every part of our daily lives.

Alexander Huth, assistant professor of neuroscience and computer science at the University of Texas at Austin, told CNN, “so, we don’t like to use the term mind reading. We think it conjures up things that we’re actually not capable of.”

Professor Huth participated in research and spent spending 20 hours in the confines of an fMRI (functional magnetic resonance imaging)  machine with audio clips he was listening to. Meanwhile, the machine captured detailed snaps of his brain activity.

OpenAIs ChatGPT logo is seen in this illustration. — Reuters/File
OpenAI’s ChatGPT logo is seen in this illustration. — Reuters/File

The AI system analysed his brain activity and the audio he was listening leading the technology to eventually foretell the words he was hearing just by watching his brain.

The technology researchers used was OpenAI’s chatGPT-1 model — which developed a huge database of books and websites.

The researchers found that the AI system accurately predicted what participants were listening to and watching by observing mental activity.

Despite its initial stages, the technology shows promise. It also underlines that AI cannot easily read our minds.

“The real potential application of this is in helping people who are unable to communicate,” Huth explained.

The researchers believed that this technology could be used in the future by people with “locked-in” syndrome, stroke and others whose brains are working but they could not speak.

“Ours is the first demonstration that we can get this level of accuracy without brain surgery. So we think that this is kind of step one along this road to actually helping people who are unable to speak without them needing to get neurosurgery,” he said.

A screen can be seen showing the OpenAI logo with ChatGPT visible behind the phone. — AFP/File
A screen can be seen showing the OpenAI logo with ChatGPT visible behind the phone. — AFP/File

Though the results of the technology are promising, it also raised concerns about how it would be used in controversial areas.

The researchers noted that brain scans “need to occur in an fMRI machine, the AI technology needs to be trained on an individual’s brain for many hours, and subjects need to give their consent.”

If someone resists listening to audio and does not think as per the requirement, it simply won’t work.

Jerry Tang, the lead author of a paper explained: “We think that everyone’s brain data should be kept private. Our brains are kind of one of the final frontiers of our privacy.”

Tang explained that “obviously there are concerns that brain decoding technology could be used in dangerous ways.”

Huth stated: “What we can get is the big ideas that you’re thinking about. The story that somebody is telling you, if you’re trying to tell a story inside your head, we can kind of get at that as well.”

Voicing concerns, Tang told CNN that lawmakers need to take “mental privacy” seriously to protect “brain data” — our thoughts — two of the more dystopian terms I’ve heard in the era of AI.

“It’s important not to get a false sense of security and think that things will be this way forever,” Tang warned.

“Technology can improve and that could change how well we can decode and change whether decoders require a person’s cooperation.”

Latest News

Facebook Content Monetization Program is launched by Meta

Published

on

By

Meta has introduced a new monetization program designed to streamline and enhance income prospects for content providers on its platform.

The Facebook Content Monetization beta, unveiled on October 2, 2024, consolidates three pre-existing monetization schemes into a unified, efficient framework.

Meta, the parent company of Facebook, has announced that the new program integrates In-stream advertisements, advertisements on Reels, and the Performance Bonus into a singular revenue structure. This unification seeks to facilitate money generation for producers across many content formats, including Reels, extended films, photographs, and textual posts.

The news is part of Facebook’s continuous initiatives to assist content creators on its platform. Meta indicates that since the launch of Facebook-funded monetization in 2017, over four million content producers have generated revenue using the platform.

Substantial Increase in Creator Compensation

Meta revealed that over the past year, Facebook has compensated content providers in excess of $2 billion for videos, reels, photographs, and text posts. Payouts for reels and short videos have significantly increased, rising by almost 80% during this timeframe.

Streamlining the Monetization Procedure

The newly introduced Facebook Content Monetization tool tackles a significant difficulty encountered by authors. Historically, diverse monetization programs had discrepancies in availability, qualifying criteria, and enrollment procedures. This intricacy led to certain creators missing possibilities or being ineligible to profit from all available forms.

Meta’s data indicates that around one-third of monetized authors on Facebook benefit from multiple Facebook-funded programs. The newly unified program seeks to augment this figure by streamlining the process and broadening income prospects.

Mechanism of the New Program

Within the Facebook Content Monetization test, creators will require participation in a single program to monetize various content forms. The application will deliver a cohesive array of analytics for monitoring performance across various content categories.

Meta asserts that the compensation structure for the new program is consistent with the current performance-based frameworks employed in Ads on Reels, In-Stream Ads, and the Performance Bonus. Compensation will remain linked to the performance of qualifying content.

Eligibility and Implementation

The preliminary stage of the beta program is restricted to invited participants only. Meta has been dispatching invitations to one million creators currently generating revenue on Facebook. The corporation intends to persist in dispatching invites in the upcoming months.

Although open registration for the program is anticipated in 2025, Meta is providing creators the chance to indicate their interest in obtaining an early access invitation to the beta. Creators seeking further information and wishing to express their interest may do so using the Facebook for Creators platform.

Effect on Current Programs

Meta has announced that the Facebook Content Monetization initiative will ultimately supplant the existing Ads on Reels, In-Stream Ads, and Performance Bonus programs. The changeover is anticipated to take place in 2025. Participation in the new initiative is voluntary for invited creators during the beta period.

Eligibility for Content

The new program enables the monetization of eligible public films, reels, photographs, and text posts. Creators and their content must follow to Meta’s regulations, including Facebook’s Partner Monetization regulations and Content Monetization Policies, to qualify for monetization.

Enhancing Opportunities for Creators

The launch of Facebook Content Monetization marks a substantial advancement in Meta’s initiatives to assist content creators on its platform. Facebook seeks to attract and maintain a varied array of producers by streamlining the monetization process and broadening revenue opportunities across various content forms.

Continue Reading

Latest News

A fresh author copyright lawsuit concerning AI training has hit Meta.

Published

on

By

Novelist Christopher Farnsworth has initiated a proposed class-action copyright lawsuit against Meta Platforms (META.O), alleging that the technology company has improperly utilized his works, among with others, to train its Llama artificial intelligence large language model.

Farnsworth asserted in the complaint filed on Tuesday that Meta provided Llama, the technology behind its AI chatbots, with thousands of pirated books to train its response to human queries.

Other authors, including as Ta-Nehisi Coates, former Arkansas governor Mike Huckabee, and comic Sarah Silverman, have filed analogous class-action lawsuits against Meta in the same court about its purported utilization of their novels for AI training.

Farnsworth’s case, initiated by the prominent class action plaintiffs’ law firm Lieff Cabraser Heimann & Bernstein, follows a federal judge’s admonition of the lead attorney in the prior case and permits distinguished litigator David Boies and additional attorneys from his firm Boies Schiller Flexner to join the plaintiffs’ team.

Representatives for Meta and Farnsworth’s legal counsel at Lieff Cabraser did not promptly reply to inquiries regarding the recent complaint on Wednesday.

Multiple collectives of copyright holders, including authors, visual artists, and music publishers, have initiated legal action against prominent technology firms for the illicit utilization of their creations in the training of generative AI systems.

The businesses contend that their AI training is safeguarded by the copyright doctrine of fair use and that the lawsuits jeopardize the emerging AI industry.

Farnsworth, a resident of Los Angeles, asserted in his lawsuit that his creations were incorporated in a “trove” of illicitly obtained literature utilized by Meta to educate Llama.

He requested the court for an indeterminate sum of monetary damages and an injunction compelling Meta to cease its purported infringement.

The case is Farnsworth v. Meta Platforms Inc, U.S. District Court for the Northern District of California, No. 3:24-cv-06893.

For Farnsworth: Elizabeth Cabraser, Daniel Hutchinson, Reilly Stoler, and Rachel Geman of Lieff Cabraser Heimann & Bernstein

Continue Reading

Latest News

PTA Chairman: I can now reveal who banned ‘X’ and ordered Pakistani internet shutdowns.

Published

on

By

Major General (Rtd) Hafeezur Rehman, Chairman of the Pakistan Telecommunication Authority (PTA), made it clear that the federal government is the only entity with the authority to stop the social media site “X,” which was formerly known as Twitter.

Informal interviews with reporters revealed that “X” will reopen as soon as the government issued the order. The Sindh High Court is currently hearing the platform’s case because of the large number of complaints it has received in Pakistan.

In response to queries on internet outages, the Chairman pointed out that Pakistan saw only seven internet outages in 2023, compared to 116 in India. He made it clear that national security still comes first even though the PTA opposes internet shutdowns.

The Chairman further stated that although internet connections were only momentarily interrupted in some locations at specific times, mobile services were suspended in certain areas on Ashura, the 10th of Muharram.

Making similarities, he pointed out that Bangladesh similarly cut off cell service during elections, adding that each nation faces different security issues and that decisions are based on intelligence gatherings.

Additionally, he reaffirmed that Panjgur’s mobile data services are temporarily suspended because of persistent security concerns.

Continue Reading

Trending