Do AI Models Hallucinate? The Answer In This New Study Might Surprise You

The launch of AI models keeps increasing with each passing day. From Google’s Gemini to OpenAI’s GPT-4o, the list is on the rise.

However, a new study is sharing some interesting findings that put to end a debate about AI models hallucinating. The conclusive answer is yes, and they’re all guilty of doing so. Some are involved so much that the end result is hilarious while in other cases, it’s a huge issue.

Not every model will make up false narratives at the same pace as the next but they’re all involved. The type of misinformation spouted depends on the kind of source involved and what it’s exposed to.

The research study from Cornell spoke about universities benchmarking the degree of hallucination with assistance from fact-checkers. As per the report, no model did great, no matter what topic was involved. Strikingly, all the models hallucinated and those that did the least answered little to no questions. Their fear of getting things wrong was so real.

As per the authors, the biggest take-home message is that AI models cannot be trusted. Even those deemed to be the best produce text that’s free from hallucinations, just 35% of the entire time.

To make the study more interesting and raise the standard of benchmarks, the authors used topics that even Wikipedia failed to have references for. More than one dozen different models were studied, and most of them were new releases.

From Meta and Google to Mistral and OpenAI, the list included them all. And as per the study, no model is hallucinating any less than before, including the biggest names in the industry.

For instance, GPT-4o didn’t perform any better in terms of fewer hallucinations than its much older counterpart GPT-3.5. Still, the AI giant had the least amount of hallucinations calculated in general, followed by Mixtral and Perplexity Sonar.

In terms of which topics proved to be the most difficult here, it was finance and the world of entertainment/celebrities. Interestingly, models found IT and geography the easiest to respond to. The logic behind this could be more training done in these domains. Where Wikipedia answers or references couldn’t be found, the models struggled a lot. This shows how most of them were trained using Wikipedia the most.

Those models who are capable of searching the web also struggled with non-Wiki queries. Did model size matter? Not really as small designs offered the same type or levels as their bigger-sized counterparts.

It’s a clear message to the world that vendors aren’t delivering what they’re claiming. The models fall short in terms of promises made or perhaps they were exaggerated too much. Whatever the case may be, it’s safe to conclude as per the authors that hallucinations and AI are in synchronizing and won’t be going away anytime soon.

What’s the solution? More policies and greater regulation seem to be the only way out, as per the researchers.

Image: DIW-Aigen

Read next: A New Research Shows that Usage of Tablets Among Children can Develop Anger Issues in Them
by Dr. Hura Anwar via Digital Information World

Eric Schmidt's Controversial Video on Google’s Work-Life Balance Disappears

A video posted by Google’s former CEO speaking negatively about the Android maker’s remote work practices has mysteriously disappeared.

Eric Schmidt did raise eyebrows when he spoke negatively about the company to students at Standford. As per Schmidt, Google could care less about winning and more about ensuring its employees had the perfect work-life balance. Even if that cost the company millions, it could care less.

The video was published across several social media platforms before but now has disappeared. Even on YouTube, its settings were changed from public to private. As expected, the matter remains a mystery in terms of why the sudden change.

The hot views were not for the masses to benefit from. Schmidt probably realized this a little too late and hence now felt it needs to be restricted. He simply expressed regret for the decisions he was making.

Many news outlets already published this with eye-grabbing headlines. Schmidt is trying hard to rectify his mistake by issuing a public apology for how his words may misrepresent the reality of Google and its working hours.

For a clip to get millions of views in such a short period and then be deleted must be a serious concern. And this is what’s happening right now.

In another part of the clip, the ex-Google CEO requests audience members to stay hush about what he’s telling them. But when someone highlights the presence of cameras, he remains startled.

There is another part in the clip where Schmidt informs some shocking claims about AI. The technology can steal IP addresses and once that’s done, their owners could hire top lawyers to clean up the mess afterwards so nothing really goes noticed as the goal is a successful product.

Who knew that his words would come back to haunt him like they are right now? Clearly, it’s a foot-in-the-mouth situation and one that Google might not forgive and forget for a long time. It was online for so long and now that it’s gone, the matter is looked upon as suspicious.

Read next: Meta Shuts Down Fact-Checking Tool CrowdTangle Just Three Months Before US Elections
by Dr. Hura Anwar via Digital Information World

Meta Shuts Down Fact-Checking Tool CrowdTangle Just Three Months Before US Elections

We are getting closer to the US elections, and that means social media apps are stepping up to the plate to curb the spread of misinformation. However, Meta is leaving many surprised with its recent decision.

Facebook’s parent firm just announced it’s getting rid of its CrowdTangle tool which not a lot of people have heard about in the past. It’s mostly used for investigating violence and false narratives in politics across its apps.

The news raised eyebrows because some felt it was a great tool for monitoring and reporting hate speech. After this, the company was bombarded with requests to rethink the call and put an end to the plans or perhaps wait until the elections were over.

So many regulators also spoke about the risks this decision brings because it’s so helpful for researchers looking for security threats and misinformation during this period.

The tool in question gave experts more insights into the company’s algorithms. People were shown details about how the data went viral. But as time went on, the tool became more and more limited in usage and accepting new users.

For now, Meta’s reps are refusing to comment on the subject. However, some reports spoke about another tool dubbed Meta Content Library. It was described as more detailed and useful that depicted the real picture better.

In that context, Nick Clegg who heads the company’s global affairs mentioned that this new tool was launched to assist people. Hence, users must use it with good intentions so everyone can benefit.

To use it, the company requires researchers with specific backgrounds to apply so it’s less hands-on in terms of access. From now on, we’re hearing more about how hundreds are already claiming it to be non-sufficient. Some do welcome it but to fill in the footsteps of CrowdTangle is going to be a challenge for many.

After the launch, Meta offered its CrowdTangle to all leading global publishers. It was interesting to see Facebook give it for free so that media partners could benefit. Clearly, it was the first time a leading social media network wished to give the public a tool for monitoring.

Many called it the most useful offering that not only identified real from fake but could also track false content. So many conspiracies were debunked including those linked to COVID, Russian operations, and the QAnon conspiracies.

Over so many years, many used it to get more insights about controversial topics. The usage was varied and so diverse. As published by some media outlets, seeing it go is going to be hard for many people.

Researchers, academicians, regulators, and beyond are mourning its loss as they fear nothing can compare to the usefulness that this provided.

Some refer to the shuttering as a huge blow to Meta and its goal to offer transparency through its apps.

Image: DIW-aigen

Read next:

• A New Research Shows that Usage of Tablets Among Children can Develop Anger Issues in Them

• YouTube Tests AI-Generated Live Chat Summaries And Super Likes

• Now Freelancers Can Receive PayPal Payments Directly from US Businesses in Their Payoneer Account
by Dr. Hura Anwar via Digital Information World

A New Research Shows that Usage of Tablets Among Children can Develop Anger Issues in Them

According to new research, giving a tablet to a child to watch videos can result in anger issues later on in his life. This is because mobile and tablets are creating problems in children’s development and are shaping them according to what they watch online. Census Bureau reported that 80% of the children own tablets in US households. The research published in JAMA Pediatrics shows that tablet use at 3.5 years old can turn into anger and frustration after a year.

4.5 years old children using tablets showed signs of outbursts and frustrations when they turned 5.5 years old and this cycle continued to upcoming years. It indicates that usage of tablets can make it harder for children to manage their emotions during their daily lives. The research didn't mention why use of tablets can make children react that way but there are two ways a child learns emotions when he is little. One is through his parents and the other is through observation.

When there is a tablet in a child’s hand he wouldn't be able to learn emotions and how to express them, which results in poor development and growth of children. It also depends on how a tablet is being used. If it is being used to read or if it is being used for watching videos, and how much parents interact with their children while they are using tablets.

The researchers also added that tablets can help children respond to emotions and behavioral cues better if a parent is interacting with the child when he's using the tablet. The research also shed some light on how usage of mobiles and tablets at such early ages can have serious negative effects on children so parents should be extra cautious of their child’s screen time.

Image: DIW-Aigen

Read next: China and the U.S. Drive 42% of Global Tech Spending, Forrester Research Reveals
by Arooj Ahmed via Digital Information World

Google Launches Gemini Live For More ‘In-Depth’ Voice Chats

Google is giving tough competition to OpenAI’s latest Advanced Voice Mode on ChatGPT with the launch of its Gemini Live.

The latest rollout is said to cater to more in-depth voice conversations on the Gemini chatbot. The product was first unveiled after getting announced at the company’s I/O developer conference this year and it’s finally going live.

For now, it’s just a limited alpha but the idea of having comprehensive voice chats through the company’s AI voice assistant is another breakthrough launch. Google promises consistent chats that are more emotionally expressive and realistic in nature. You can also speak in between the chatbot’s conversation by interrupting and making the most of multi-turn dialogue as well.

Before, asking follow-up queries was not possible. Now, the chatbot is enabled to follow your tone and pattern of speech live.

Google described the entire matter through a new blog post where it boasted the 10 different natural-sounding voices that enable you to provide responses. Hence, no need to worry about your tone, speed, or interruptions as Gemini Live will adjust accordingly. Just pretend like you’re talking to a human being as everything stays the same.

Other than that, Google says it’s hands-free, an option that those on the run can avail. Keep on talking into the app or when the device is locked. All the chats could be paused and then resumed at any given moment in time.

As far as productivity is concerned, it’s right up there. Imagine preparing for a job interview. You can simply rehearse before the next meeting or practice giving a speech to a chatbot with ease. The produce allows you to gain more insights on how to make your speaking skills better, including some useful tips and tricks to nail any conversation that puts you on the spot.

Another major advantage over OpenAI’s Advanced Voice Mode is enhanced memory. The offering has a longer context window. This means taking in a lot of information and providing logical reasoning regarding that. So you can go on and on, and engage in back-and-forth discussions, before finally producing a written response.

Google says Gemini Live makes use of the newest models that give it a more conversational touch. It’s a huge context window to benefit from where users can take part in long chats through Live.

The promises are plenty but how well Google stands against them is a question worth pondering that only time can tell. There is one thing missing that Google displayed at the developer’s conference for this product. It’s the multimodal output.

Google first rolled out pre-recorded content displaying Gemini Live showing responses to users’ backgrounds through images and recordings captured via phone cameras. Hence, Google did speak about how it’s arriving sometime this year. But it failed in terms of adding more details on the subject.

Gemini Live will also be rolled out in different languages this year and can also be available for iOS users. For now, it’s just in English. Also, it’s not free. Only those with Gemini Advanced subscriptions will benefit which costs $20 a month.

However, Google does promise more freebies for Gemini coming soon this year. This includes making overlay more prominent on the app through more questions about what’s next. Simply press the Power button and say the term ‘Hey Google’ to produce a vast array of images. These can be edited into other Google apps such as Gmail.

Google is also equipping Gemini with more Google services. This includes more activities with Calendar, YouTube Music, Tasks, and Utilities. Media, volume, Bluetooth, flashlight, and more.

As per the company’s latest blog post, it looks like there’s a lot you can do with Gemini. The list of tasks is plenty and it’s like your own mini assistant. From finding what to cook for dinner to curating the perfect music playlist for date night, trust Gemini to do it all.

Did we mention how Gemini is coming to Android tablets soon? Clearly, there’s a lot to anticipate here and we’ll keep you updated as it happens.

Read next:

• China and the U.S. Drive 42% of Global Tech Spending, Forrester Research Reveals

• FCC Intensifies Fight Against AI-Generated Phone Calls With New Rules

• YouTube Capitalizes On Shorts Popularity By Testing Longer Clips Within The Shorts Feed

• What Are The Industrial Applications Of Nanoparticle Processing?
by Dr. Hura Anwar via Digital Information World

What Are The Industrial Applications Of Nanoparticle Processing?

Nanoparticle processing is a rapidly growing field that involves the manipulation and production of nanoparticles for various industrial applications. These tiny particles, typically measuring between 1 and 100 nanometers, exhibit unique properties that make them invaluable in a wide range of industries. From improving material strength to enhancing drug delivery systems, nanoparticle processing is revolutionizing the way we approach many technological and scientific challenges. This article explores the diverse industrial applications of nanoparticle processing and its impact on various sectors.

Nanoparticle Processing In Medicine

Targeted Drug Delivery

One of the most significant applications of nanoparticle processing in the medical field is in the development of targeted drug delivery systems. Nanoparticles can be engineered to carry therapeutic agents directly to specific cells or tissues, improving the efficacy of treatments while minimizing side effects. This precision targeting is particularly beneficial in the treatment of cancer, where nanoparticles can deliver chemotherapy drugs directly to tumor cells, sparing healthy tissue.

Image: DIW-Aigen

Nanoparticle-based drug delivery systems also enhance the solubility and stability of drugs, making it possible to develop formulations for drugs that were previously difficult to administer. This innovation has led to the creation of more effective treatments for a variety of diseases, from cancer to cardiovascular conditions.

For more information on the advancements in nanoparticle-based drug delivery, you can refer to Science Daily.

Diagnostics And Imaging

Nanoparticles are also playing a crucial role in medical diagnostics and imaging. Gold nanoparticles, for instance, are used in diagnostic assays to detect biomarkers associated with various diseases, including cancer and infectious diseases. These assays provide high sensitivity and specificity, enabling early and accurate diagnosis.

In medical imaging, nanoparticles are used to enhance the contrast of images, allowing for better visualization of tissues and organs. For example, iron oxide nanoparticles are used as contrast agents in magnetic resonance imaging (MRI) to improve the clarity and detail of the images. This advancement has significant implications for the early detection and treatment of diseases.

Nanoparticle Processing In Electronics

Improved Electronic Components

Nanoparticle processing has had a transformative impact on the electronics industry, particularly in the development of advanced electronic components. Quantum dots, which are semiconductor nanoparticles, are used in the production of high-performance display technologies, such as quantum dot displays (QLEDs). These displays offer superior brightness, color accuracy, and energy efficiency compared to traditional display technologies.

Nanoparticles are also used in the fabrication of thin-film transistors and memory devices, where they provide enhanced electrical properties and greater miniaturization. This has enabled the production of smaller, faster, and more efficient electronic devices, from smartphones to computers.

Energy Storage And Conversion

In addition to improving electronic components, nanoparticle processing is critical in the development of advanced energy storage and conversion technologies. Nanostructured materials are used to create batteries with higher energy densities and longer lifespans. For instance, lithium-ion batteries with nanoparticle-enhanced electrodes offer greater storage capacity and faster charging times.

Nanoparticles are also used in the production of more efficient solar cells. Perovskite nanoparticles are being explored for their potential to increase the efficiency of solar energy conversion, making renewable energy sources more viable and cost-effective.

For updates on the latest innovations in nanoelectronics and energy technologies, check out TechCrunch.

Nanoparticle Processing In Materials Science

Enhanced Material Properties

Nanoparticle processing is revolutionizing materials science by enabling the development of materials with enhanced properties. By incorporating nanoparticles into materials, researchers can significantly improve their mechanical, thermal, and electrical characteristics.

Carbon nanotubes and graphene are prime examples of nanomaterials that have been used to create composite materials with extraordinary strength and conductivity. These materials are being used in a variety of applications, from aerospace components to sports equipment, where they provide lightweight and durable solutions.

Coatings And Surface Treatments

Nanoparticles are also used to develop advanced coatings and surface treatments that enhance the properties of various materials. Nanocomposite coatings can provide surfaces with improved hardness, corrosion resistance, and self-cleaning capabilities. These coatings are widely used in industries such as automotive, aerospace, and construction to extend the lifespan and performance of materials.

In addition, antimicrobial coatings containing nanoparticles are being developed to prevent the growth of bacteria and other microorganisms on surfaces. This has important applications in healthcare and food processing, where maintaining hygienic conditions is crucial.

Nanoparticle Processing In Environmental Applications

Pollution Control And Remediation

Nanoparticle processing is playing a vital role in environmental protection through the development of technologies for pollution control and remediation. Nanoporous materials and nanocatalysts are used to remove pollutants from air and water, offering more efficient and effective solutions compared to traditional methods.

For example, nanoparticles are used in water treatment systems to remove heavy metals, organic pollutants, and pathogens, providing access to clean water in areas where resources are limited. In air pollution control, nanoparticles are used in filters and catalysts to reduce emissions from industrial processes and vehicles.

Sustainable Energy Solutions

Nanoparticle processing is also contributing to the development of sustainable energy solutions. Nanomaterials are used in the production of more efficient solar cells, batteries, and fuel cells, which are essential for reducing our reliance on fossil fuels and mitigating the impacts of climate change.

By improving the performance and cost-effectiveness of renewable energy technologies, nanoparticle processing is helping to drive the transition to a more sustainable energy future.

To explore more about the environmental applications of nanotechnology, visit Nature.

Industrial Applications In Manufacturing

Additive Manufacturing

Nanoparticle processing is revolutionizing additive manufacturing, also known as 3D printing. Nanoparticles are used to create advanced inks and filaments that enable the production of high-precision and high-performance components. These materials offer improved mechanical properties, such as strength and flexibility, and can be used to produce complex geometries that are difficult to achieve with traditional manufacturing methods.

Catalysis And Chemical Processing

In the chemical industry, nanoparticles are used as catalysts to enhance the efficiency of chemical reactions. Nanocatalysts offer a larger surface area and improved reactivity compared to traditional catalysts, enabling faster and more selective chemical processes. This leads to increased production efficiency and reduced waste, making chemical manufacturing more sustainable.

Nanoparticle-Based Lubricants

Nanoparticle processing is also used to develop advanced lubricants that reduce friction and wear in industrial machinery. Nanoparticle-based lubricants offer superior performance compared to conventional lubricants, extending the lifespan of equipment and reducing maintenance costs. These lubricants are used in a variety of industries, including automotive, aerospace, and manufacturing.

Applications By Nanoparticle Processing Companies

Nanoparticle processing companies are driving innovations across various industries, contributing to advancements in material science, electronics, and medicine. Companies like DIANT Pharma specialize in developing and applying nanoparticle technology for industrial applications. DIANT Pharma’s expertise in nanoparticle processing enables them to provide high-quality materials that meet the specific needs of their clients, ranging from pharmaceuticals to environmental solutions.

DIANT Pharma: Innovating Industrial Applications

DIANT Pharma is a leading company in the field of nanoparticle processing, offering innovative solutions for various industrial applications. DIANT Pharma specializes in the development and production of nanoparticles for use in pharmaceuticals, electronics, and environmental technologies. Their expertise in nanoparticle processing enables them to provide high-quality materials that meet the specific needs of their clients.

For more information on DIANT Pharma’s contributions to nanoparticle processing, you can explore their latest developments at nanoparticle processing.

BASF: Advancing Nanotechnology Solutions

BASF is a global leader in the development and application of nanotechnology solutions for various industries. The company offers a wide range of nanoparticle-based products, including advanced materials, coatings, and catalysts. BASF’s commitment to innovation and sustainability has made it a trusted partner for companies seeking to leverage the benefits of nanotechnology in their operations.

For updates on BASF’s advancements in nanotechnology, visit Science Daily.

Future Directions In Nanoparticle Processing

Integration With Digital Technologies

The integration of digital technologies such as artificial intelligence (AI) and machine learning with nanoparticle processing is set to revolutionize the field. AI algorithms can analyze large datasets generated during nanoparticle synthesis and processing, optimizing production parameters and improving product quality.

Machine learning models can also predict the properties and performance of nanoparticles based on their composition and processing conditions, enabling the development of tailored materials for specific applications. The combination of digital technologies and nanoparticle processing will drive further innovation and enhance the capabilities of various industries.

Sustainable Nanomanufacturing

As the demand for nanoparticle-based products continues to grow, there is an increasing focus on developing sustainable nanomanufacturing processes. This includes the use of environmentally friendly solvents and materials, as well as the implementation of energy-efficient production methods. Sustainable nanomanufacturing not only reduces the environmental impact of nanoparticle production but also contributes to cost savings and improved operational efficiency.

Advanced Characterization Techniques

Advancements in characterization techniques are also set to enhance the precision and accuracy of nanoparticle processing. Techniques such as high-resolution electron microscopy and spectroscopy enable detailed analysis of nanoparticle size, shape, and composition, providing valuable insights into their properties and behavior. These advancements will support the development of more effective and reliable nanoparticle-based products.

by Asim BN via Digital Information World

China and the U.S. Drive 42% of Global Tech Spending, Forrester Research Reveals

Forrester Research is shedding light on the world’s top six ranked digital economies by size. The global digital economy is expected to grow at a 7% compound annual growth rate (CAGR) from $11.8 trillion in 2023 to $16.5 trillion in 2028.

The American digital economy will see a 6.4% growth rate, reaching $6 trillion by 2028. Gains from digital apps are not equally distributed. Major players like Microsoft, Amazon, Oracle, and SAP captured 39% of the software market share last year.

Forrester’s estimates are based on financial documents, executive interviews, and internet traffic analysis. The main growth drivers are online retail and online travel, with expected CAGR rates of 9% and 7%, respectively, from 2023 to 2028.

China and the U.S. drive most of the global digital economy, accounting for 42% of global tech spending, despite having only 4.2% of the world’s population and 26% of global GDP share. France, Brazil, Canada, Japan, Germany, the U.S., the U.K., Spain, and Italy are the largest contributors to local digital economies due to their significant technology spending.

By 2028, the digital economy will account for approximately 31% of South Korea’s GDP. This reflects South Korea's large and balanced digital sector. In contrast, Mexico's digital economy will contribute about 10% to GDP, and Brazil's will contribute 8%. South Korea benefits from high consumer and business spending, with room for further growth in government and enterprise tech spending in China and consumer digital spending in Australia.

The report, Global Digital Economy Forecast, 2023-28, measures the size and growth of the digital economy across tech spend, retail, travel ecommerce, insurance ecommerce, ICT exports, chip manufacturing, and data centers.

Global Digital Economy Forecast	2022 ($ billions)	2028 Forecast ($ billions)
United States	$4139 Billion	$6006 Billion
China	$3104 Billion	$4700 Billion
United Kingdom	$615 Billion	$852 Billion
Japan	$499 Billion	$641 Billion
Germany	$511 Billion	$750 Billion
South Korea	$442 Billion	$648 Billion
France	$414 Billion	$607 Billion
India	$329 Billion	$736 Billion
Canada	$271 Billion	$371 Billion
Italy	$181 Billion	$272 Billion
Spain	$168 Billion	$262 Billion
Australia	$146 Billion	$195 Billion
Mexico	$147 Billion	$228 Billion
Brazil	$139 Billion	$224 Billion

Read next: The Global Winners and Losers: Most Loved and Hated Email, Messaging, and Social Media Apps Revealed
by Dr. Hura Anwar via Digital Information World