Metaschief AI scientist thinks large language models will never reach human intelligence.

Yann LeCunasserts that artificial intelligence (AI) large language models (LLMs) such as ChatGPT have alimited grasp on logic, the Financial Times (FT) reported Wednesday (May 21).

These models, LeCun told the FT, do not understand the physical world, do not have persistent memory, cannot reason in any reasonable definition of the term and cannot plan...hierarchically.

He argued against depending on LLMs to reach human-level intelligence, as these models need the right training data to answer prompts correctly, thus making them intrinsically unsafe.

LeCun is instead working on a totally new cohort of AI systems that aim to power machines with human-level intelligence, though this could take 10 years to achieve.

The report notes that this is a potentially risky gamble, as many investors are hoping for quick returns on their AI investments. Meta recently saw its value shrink by almost $200 billion after CEO Mark Zuckerbergpledged to up spendingand turn the tech giant into the leading AI company in the world.

Meanwhile, other companies are moving forward with enhanced LLMs in hopes of creating artificial general intelligence (AGI), or machines whose cognition surpasses humans.

For example, this week saw AI firmScaleraise $1 billion in a Series F funding round that valued the startupat close to $14 billion, with founder Alexandr Wang discussing the companys AGI ambitions in the announcement.

Hours later, the French startup called H revealed it had raised $220 million, with CEO Charles Kantor telling Bloomberg News the company is working towardfull-AGI.

However, some experts question AIs ability to think like humans. Among them isAkli Adjaoute, who has spent 30 years in the AI field and recently authored the book Inside AI.

Rather than speculating about whether the technology willthink and reason, he views AIs role as an effective tool, stressing the importance of understanding AIs roots in data and its limitations in replicating human intelligence.

AI does not have theability to understandthe way that humans understand, Adjaoute told PYMNTS CEO Karen Webster.

It follows patterns. As humans, we look for patterns. For example, when I recognize the number 8, I dont see two circles. I see one. I dont need any extra power or cognition. Thats what AI is based on. Its the recognition of algorithms and thats why theyre designed for specific tasks.

Read more:
Meta AI Head: ChatGPT Will Never Reach Human Intelligence - PYMNTS.com

In a new interview with Elena Rosberg of Radiocast BG, JUDAS PRIEST bassist Ian Hill was asked how he thinks heavy metal music and the metal community can combat the negative effects of artificial intelligence in music, particularly as it relates to the creative process. Hill responded (as transcribed by BLABBERMOUTH.NET): "I don't know. You'd know, wouldn't you? If something was put together they're not gonna be able to play the instruments for a start, so they're gonna have to use some kind of music stem, some sort of source. And you'd know. I'd know, I'd know if something was put together. You can listen to I don't know even now songs, pop songs in particular, and you think, 'That's not a bass guitar. Some guy's playing that on a keyboard.' You know it. It might fool some of the public, because, obviously, I go into things a little bit more deeply. I drive my wife mad. She won't let me anywhere near If she's going to a concert, I don't go, basically because I start picking it apart. But that's what we do. A welder would do the same thing. He'd look at something [and go], "That's a load of rubbish.' That's what you do."

He continued: "But I don't know. I just think that artificial intelligence can't really perform live. I mean, this is what it's coming down to. A lot of music, especially in the pop world these days, is a little bit on the false side people mime to it and what have you. And A.I. can't even do that. You can't have artificial intelligence standing on stage. That ain't gonna work. So, from a recording point of view, yeah, they might fool people they might fool a hell of a lot of people but, actually, when they say there's a band playing live, that's gonna be the acid test, isn't it? And I can't see, really, unless they're all holograms standing up there. which has been done. What am I saying? ABBA have just done it, haven't they? But it's there. It is advertised. You know it ain't ABBA. It's trickery. But it's in the live performance where it'll fall down and it won't stand up to scrutiny, I don't think."

JUDAS PRIEST kicked off the U.S. leg of the "Invincible Shield" world tour on April 18 at Toyota Oakdale Theatre in Wallingford, Connecticut.

Hill is the sole remaining original member of PRIEST, which formed in 1969. Singer Rob Halford joined the group in 1973 and guitarist Glenn Tipton signed on in 1974. Rob left PRIEST in the early 1990s to form his own band, then came back to PRIEST in 2003. Original guitarist K.K. Downing parted ways with the band in 2011, and was replaced by Richie Faulkner.

PRIEST's latest album, "Invincible Shield", entered the U.K. chart at No. 2, just behind Ariana Grande's "Eternal Sunshine".

Prior to "Invincible Shield"'s arrival, PRIEST's highest U.K. chart achievement was with 1980's "British Steel", which reached No. 4.

PRIEST's 2018 album "Firepower" entered the chart at No. 5.

"Invincible Shield" was JUDAS PRIEST's fifth Top 10 album, after the aforementioned "British Steel" and "Firepower", as well as 2014's "Redeemer Of Souls" (No. 6) and the 1979 live album "Unleashed In The East" (No. 10).

"Invincible Shield" landed at No. 1 in Germany, Finland, Sweden and Switzerland, as well as No. 5 in France, No. 8 in Italy and No. 16 in Australia.

Read more from the original source:
JUDAS PRIEST's IAN HILL Weighs In On Use Of Artificial Intelligence In Music - BLABBERMOUTH.NET

MORGANTOWN, W.Va. Researchers from West Virginia University, Marshall University, and Florida International University are exploring the cybersecurity needs of artificial intelligence technologies with a $1.75 million grant from the Defense Advanced Research Projects Agency (DARPA).

Professor and Chairman of the Statler College of Engineering and Mineral Resources in the Lane Department of Computer Science and Electrical Engineering, Anurag Srivastava, said the AI-CRAFT project is intended to develop ways to secure the emerging technology. Artificial intelligence is developing rapidly and being pushed into larger real-time applications.

Our goal is to look at what those are, and how do I defend myself if someone is trying to hack into or make the AI behave in a way it is not supposed to behave? Srivastava said.

The teams are building the artificial intelligence systems while engineering security and safety as they are deployed. The complexity results from the evolution from simple calculation systems used a decade ago to the addition of millions of data points as AI systems are taught to think like the human brain.

Look at this from a new point of view now; what is the attack vector now? Srivastava asked. Can someone reverse engineer the AI? Can someone poison the AI so it behaves in a way it should not?

AI technologies have quickly grown from application in autonomous cars to use in very lifelike robots and vital systems like public utilities. The research also includes developing secure data practices, access controls, and continuous monitoring to assess the security and usefulness of AI systems.

Especially those that will be used to operate critical systems like robots or the power grid, Srivastava said.

On the academic side, students will have many hands-on opportunities in labs and training platforms, designed to equip students with the skills and knowledge needed to thrive in a rapidly evolving cybersecurity industry.

Other than solving this problem for complex systems like a power grid, robotics, or autonomous cars, our goal is also to teach it because this is also a new topic, Srivastava said.

Officials from WVU, Marshall University, and the U.S. Department of Defense will break ground on the new Institute for Cyber Security in Huntington on May 17.

Go here to read the rest:
Study including WVU and Marshall analyzes cyber threats to Artificial Intelligence systems - West Virginia MetroNews

Investors using generalized, all-purpose AI platforms for retirement planning help may not even realize these tools' shortcomings.

Have you ever considered asking an artificial intelligence (AI)-powered assistant like Google's Gemini or Microsoft's Copilot for help allocating your retirement portfolio? If so, you wouldn't be the first. You may already be using AI to allocate your retirement savings without knowing it.

These tools are obviously powerful by virtue of being able to access much of the world's collective knowledge. A handful of brokerages and investment managers know this and have been providing such technology for some time now. They just don't tell clients that their stock picks are being powered by artificial intelligence because ... well, because the world's still not completely convinced that generalized AI-powered chat tools like OpenAI's ChatGPT or Gemini always get things right.

That's because the AI doesn't always get things right, by the way. Indeed, when it comes to retirement planning, all-encompassing platforms like the aforementioned Copilot or ChatGPT can sometimes be alarmingly problematic by missing key details about your specific situation.

With that as the backdrop, here's why you're better served by sticking with the AI tools built specifically to help you manage your retirement investments.

If you're a customer of brokerage firm Charles Schwab (SCHW 0.95%), then you've already been exposed to an AI-powered stock-picking tool. Unveiled in early 2022, this technology allows investors to narrow their investment choices down to a manageable handful based on the usual criteria like risk, growth, and valuation. But the tool also helps investors identify and capitalize on qualitative trends or themes.

It's not exactly a new idea. Investors have had access to theme-based stock suggestions for years now. Usually, it just seems like another criterion offered by a stock screener.

There's far more artificial intelligence being applied here than there seems to be on the surface. Much like Microsoft's Copilot and Google's Gemini, Schwab's tool uses AI "to find companies linked to those keywords and phrases, combing through millions of public documents, such as patents, clinical trials, and company filings." It very likely comes up with picks most investors would have otherwise never come across.

Perhaps more relevant to future or current retirees is Schwab's menu of so-called Intelligent Portfolios, which build and then update a portfolio pre-selected for your particular situation. What the company refers to as a robo-advisor doing this portfolio maintenance is actually a form of automated AI. Although these portfolios certainly don't outperform the broad market each and every quarter, they do offer less volatile, more consistent performances without requiring constant monitoring and management by investors.

Image source: Getty Images.

Mutual fund giant Fidelity also has its own robo-advisor tech, powering an automated investment service called Fidelity Go. Although it's not Fidelity's invention, the brokerage firm offers its customers access to Capitalize.ai, allowing users to turn the simplest of word-based instructions into a trading algorithm.

Investment management outfit BlackRock (BLK 0.56%) -- the name behind the iShares family of exchange-traded funds (ETFs) -- is another financial services name waist-deep in AI waters.

While most of the company's interest in AI has been to empower financial advisors via a revenue-bearing platform called Aladdin, it's not unaware of the potential of AI as a stock-picking tool. For instance, commenting specifically on retirement portfolios, BlackRock explains that "by analyzing vast datasets, including satellite imagery and labor mobility data, AI can extract early insights on economic activities across regions, which can be used to inform macro (e.g., regional) and micro (e.g., company level) tilts in our portfolios."

To this end, while it doesn't offer a Schwab-like robo-advisor to individual investors, BlackRock has never denied the fact that it's using AI to help guide its fund managers' investment choices. So, in a sense, some individual investors are directly benefiting from BlackRock's AI tech.

Some might argue that general-purpose AI chat platforms like ChatGPT and Gemini have come a long way since Schwab and Fidelity first launched their AI-powered stock-picking tools. Namely, they've become incredibly user-friendly.

And that's not incorrect.

But all-purpose AI platforms still aren't great self-service options for allocating your retirement savings even when you can convince a general AI assistant to help you do so. These platforms are seemingly aware that even they are not ideally suited to offer you a full-blown portfolio plan; as such, they don't always provide more than a broad allocation theory.

Not all AI is the same. Microsoft's Copilot and OpenAi's ChatGPT, and Google's Gemini are all large language model (or LLM) AI, meaning they rely on the collective wisdom created by the text available on billions and billions of web pages. By interpreting and integrating as much credible information available on the web as they can, these platforms can end up ignoring a nuance that might be uniquely important to you.

It's also worth adding that while LLMs can analyze numerical data, they're not built to deal with numbers. They often struggle when it comes to doing predictive numerical analysis. This, of course, is the kind of math retirees typically need done (and need done right).

The missing element? Context. Gemini or Copilot "understand" words largely based on the words appearing before or after them. With numbers though, these tools don't always understand what it is you're trying to accomplish or determine, or how to do it, particularly when you're giving these AI platforms instructions using words.

That's not the case with the portfolio-allocation tools offered by Schwab or Fidelity, or perhaps eventually by BlackRock. Their technology is built from the ground up to handle numbers and meet needs specific to investors. They may require more input from you on the front end, but they're far better suited to provide you with the solutions you need in the end.

These investor-oriented AI tools also do something important that most general AI-powered chatbots don't; they ask you questions that force you to think about what it is you ultimately want to accomplish. ChatGPT, Gemini, and Copilot don't ask any follow-up, clarifying questions. In fact, these tools don't even recognize that they should be asking more probing questions of investors using them.

Connect the dots. The additional information and investment of your time in the creation of a customized retirement portfolio will very likely lead to superior returns (relative to the risk you're taking) from your retirement fund. So, stick with the tools purpose-built to allocate a portfolio specifically for you. It's worth it even if they're not the easiest and most accessible option to use.

Suzanne Frey, an executive at Alphabet, is a member of The Motley Fools board of directors. Charles Schwab is an advertising partner of The Ascent, a Motley Fool company. James Brumley has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Charles Schwab, and Microsoft. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft, short January 2026 $405 calls on Microsoft, and short June 2024 $65 puts on Charles Schwab. The Motley Fool has a disclosure policy.

See more here:
Surprise: Artificial Intelligence (AI) Might Already Be Used to Allocate Your Retirement Savings - The Motley Fool

Artificial intelligence (AI) computer programs that process MRI results show differences in how the brains of men and women are organized at a cellular level, a new study shows. These variations were spotted in white matter, tissue primarily located in the human brains innermost layer, which fosters communication between regions.

Men and women are known to experience multiple sclerosis, autism spectrum disorder, migraines, and other brain issues at different rates and with varying symptoms. A detailed understanding of how biological sex impacts the brain is therefore viewed as a way to improve diagnostic tools and treatments. However, while brain size, shape, and weight have been explored, researchers have only a partial picture of the brains layout at the cellular level.

Led by researchers at NYU Langone Health, the new study used an AI technique called machine learning to analyze thousands of MRI brain scans from 471 men and 560 women. Results revealed that the computer programs could accurately distinguish between biological male and female brains by spotting patterns in structure and complexity that were invisible to the human eye. The findings were validated by three different AI models designed to identify biological sex using their relative strengths in either zeroing in on small portions of white matter or analyzing relationships across larger regions of the brain.

Our findings provide a clearer picture of how a living, human brain is structured, which may in turn offer new insight into how many psychiatric and neurological disorders develop and why they can present differently in men and women, said study senior author and neuroradiologist Yvonne W. Lui, MD.

Dr. Lui, a professor and vice chair for research in the Department of Radiology at NYU Grossman School of Medicine, notes that previous studies of brain microstructure have largely relied on animal models and human tissue samples. In addition, the validity of some of these past findings has been called into question for relying on statistical analyses of hand-drawn regions of interest, meaning researchers needed to make many subjective decisions about the shape, size, and location of the regions they choose. Such choices can potentially skew the results, says Dr. Lui.

The new study results, published online May 14 in the journal Scientific Reports, avoided that problem by using machine learning to analyze entire groups of images without asking the computer to inspect any specific spot, which helped to remove human biases, the authors say.

For the research, the team started by feeding AI programs existing data examples of brain scans from healthy men and women and also telling the machine programs the biological sex of each brain scan. Since these models were designed to use complex statistical and mathematical methods to get smarter over time as they accumulated more data, they eventually learned to distinguish biological sex on their own. Importantly, the programs were restricted from using overall brain size and shape to make their determinations, says Lui.

According to the results, all of the models correctly identified the sex of subject scans between 92 percent and 98 percent of the time. Several features in particular helped the machines make their determinations, including how easily and in what direction water could move through brain tissue.

These results highlight the importance of diversity when studying diseases that arise in the human brain, said study co-lead author Junbo Chen, MS, a doctoral candidate at NYU Tandon School of Engineering.

If, as has been historically the case, men are used as a standard model for various disorders, researchers may miss out on critical insight, added study co-lead author Vara Lakshmi Bayanagari, MS, a graduate research assistant at NYU Tandon School of Engineering.

Bayanagari cautions that while the AI tools could report differences in brain-cell organization, they could not reveal which sex was more likely to have which features. She adds that the study classified sex based on genetic information and only included MRIs from cisgendered men and women.

According to the authors, the team next plans to explore the development of sex-related brain structure differences over time to better understand environmental, hormonal, and social factors that could play a role in these changes.

Funding for the study was provided by the National Institutes of Health grants R01NS119767, R01NS131458, and P41EB017183, as well as by the United States Department of Defense grant W81XWH2010699.

In addition to Dr. Lui, Chen, and Bayanagari, other NYU Langone Health and NYU researchers involved in the study were Sohae Chung, PhD, and Yao Wang, PhD.

Shira Polan Phone: 212-404-4279 Shira.Polan@NYULangone.org

Read the rest here:
Artificial Intelligence Tool Detects Sex-Related Differences in Brain Structure - NYU Langone Health