SwissCognitive Guest Blogger: Eleanor Wright, COO at TelWAI – “Fortifying the Future: Ensuring Secure and Reliable AI”

As AI becomes further integrated into various domains, from infrastructure to defence, ensuring its robustness will become a matter of national security. An AI system managing power grids, security apparatus, or financial networks could present a single point of failure if compromised or manipulated. Historical incidents, such as the Stuxnet cyberweapon, illustrate the physical and cyber damage that can be inflicted. When considering AI’s complexity, the potential for a cascade of both physical and digital harm increases dramatically.

As such, we should ask: How do we fortify AI?

AI systems must be designed to withstand attacks. From decentralisation to layering, these systems should be constructed so that control points can seamlessly enter and exit the loop without disabling the broader system. Thus, building redundancy and backup at various control points within the AI systems. For example, suppose a sensor or a group of sensors is deemed to have failed or been corrupted. In that case, the broader system must be capable of automatically readjusting to stop utilising data and intelligence gathered from said sensors.

Another strategy for strengthening AI systems involves simulating data poisoning attacks and training AI systems to detect such threats. By teaching the systems to recognise and respond to attacks or failures, they can automatically reconfigure without the need for human intervention. If an AI can learn to identify tainted data, such as statistical anomalies or inconsistent patterns, it could flag or quarantine suspect inputs. This approach leans heavily on machine learning’s strengths: pattern recognition and adaptability. However, it’s not a failsafe; adversaries could evolve their attacks to more closely mimic legitimate data, so the training would need to be dynamic, constantly updating to match new threat profiles.

Maintaining a human in the loop to enable oversight and override is considered one of the most crucial elements in the rollout of AI in various industries. Allowing humans to oversee AI decision-making and restricting autonomy can prevent potentially harmful actions taken by these systems. Whilst critical in the early stages of AI deployment as capabilities scale and evolve, there may come a point where human oversight inhibits these systems and, in itself, causes more harm than good.

Finally, AI sovereignty may prove to be the most critical element in ensuring companies and governments fully control essential algorithms and hardware powering their operations. Without this control, these systems could be vulnerable to foreign interference, including cyberattacks, espionage, or sabotage. As the use of AI increases, the sovereignty of AI systems and their components will become increasingly important. At its core, AI sovereignty is about control, whether exercised by governments, corporations, or individuals. Through the control of data, infrastructure, and decision-making power, those who build and deploy AI systems and sensors gain control of AI.

Fortification will involve integrating resilience, adaptability, and sovereignty into AI’s DNA, ensuring it is not only intelligent but also resilient and unbreakable. It can provide technological advantages, but it may also expose systems to disruption and vulnerability exploitation. As organisations race to harness AI’s potential, the question looms: Will AI enable organisations to gain a strategic advantage, or will it undermine the very systems it was designed to strengthen?

About the Author:

Holding a BA in Marketing and an MSc in Business Management, Eleanor Wright has over eleven years of experience working in the surveillance sector across multiple business roles.

Der Beitrag Fortifying the Future: Ensuring Secure and Reliable AI erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Major AI Funding Shifts – SwissCognitive AI Investment Radar

The AI investment landscape continues to evolve, with new funding rounds and strategic commitments driving the industry forward. This week, key players across finance, technology, and infrastructure have made major moves to expand AI capabilities, focusing on both software and cloud expansion. Salesforce pledged $1 billion toward AI development in Singapore, while Honor committed $10 billion to integrating AI across its product line.

Investment priorities are shifting from AI chips to software, with analysts predicting that software firms will capture more value in the coming years. Microsoft is expanding its cloud and AI infrastructure in South Africa with a $298 million investment, reflecting the rising demand for AI-driven services. Meanwhile, Barclays analysts note that AI models are evolving from training-based systems to more advanced reasoning engines, signaling a new phase in AI capabilities.

DeepSeek’s breakthrough continues to drive activity in China’s venture capital sector, attracting fresh funding after years of stagnation. Elsewhere, private equity firms are adjusting their investment strategies to keep pace with AI-driven business transformations.

With AI playing a bigger role in stock markets, investor sentiment is shifting as automation takes on a larger role in financial decision-making. The rise of AI-powered fintech solutions, such as Finnomena’s partnership with Google Cloud, further highlights the increasing role of AI in investment strategies.

Stay tuned as we track these developments and more, bringing you the latest insights from the growing AI investment world.

Previous SwissCognitive AI Radar: $100B for AI Chips, $40B for AI Bets.

Our article does not offer financial advice and should not be considered a recommendation to engage in any securities or products. Investments carry the risk of decreasing in value, and investors may potentially lose a portion or all of their investment. Past performance should not be relied upon as an indicator of future results.

Der Beitrag Major AI Funding Shifts – SwissCognitive AI Investment Radar erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

SwissCognitive Guest Blogger: Eleanor Wright, COO at TelWAI – “A New Era of Intelligent Robots”

Imagine a world where humanoid robots cook for you, care for your loved ones, and streamline your workday – all powered by AI smarter than ever before. The global AI in robotics market, projected to surpass $124 Billion by 2030, is set to make this vision a reality. As the capabilities of AI evolve, these machines will become our companions, caregivers, and coworkers, they’ll make mobility more affordable, transform access to services, and redefine the value of human effort.

From Amazon’s fleet of 750,000 warehouse robots to Tesla’s ambitions to build 10,000 humanoid Optimus robots this year, the age of robots is upon us. Dependent on sensors and actuation systems to navigate and interact with the physical environment, this new age of robotics hinges on the developments of AI, designed to mimic and learn from its biological makers. Equipping these robots with intelligence, engineers working across various domains of expertise, utilise AI to enable vision, natural language processing, sound processing, pressure sensing, and more.

Beyond sensing, AI also enables robots to reason, adapt, and learn, using approaches including—but not limited to—reinforcement learning, neural networks, and Bayesian networks. These models and methods enable robots to assess risks and determine actions, and by learning from experience, robots can adapt to new tasks and environments. Thus, AI enables robots to perceive, act, learn, and adapt, allowing them to perform tasks with greater autonomy and precision.

However, integrating AI into robotics isn’t seamless, it comes with hurdles. Robots struggle with real-time processing delays, adapting to messy unpredictable environments, squeezing efficiency from limited hardware, and understanding human quirks like vague commands or gestures. These challenges constrain capabilities and the pace at which robots enter and dominate markets.

So, how can these challenges be addressed?

Some developments in addressing these challenges include:

1. Parallel computing

Parallel computing involves dividing larger tasks into smaller, independent tasks that can be processed simultaneously rather than sequentially. This enables increased computational efficiency, reduced latency, and improved cost efficiency. In robotics, parallel computing allows robots to process inputs from LIDAR, radar, and cameras simultaneously, enabling them to navigate environments more effectively and efficiently.

2. Transfer learning

Transfer learning leverages pre-trained models to solve new, but similar, problems. In this approach, a model trained on one task or dataset is reused and fine-tuned for a related task. For example, in machine vision for defect detection in manufacturing, fine-tuning a pre-trained model on a smaller dataset of images allows it to quickly adapt to detect specific defects, such as cracks or dents, without needing to train a model from scratch.

3. Self-calibrating AI

Self-calibrating refers to AI systems that autonomously adjust their parameters, models, or processes to maintain optimal performance without manual intervention. In robotics, self-calibrating AI enables robots to adapt to changes in their environment, hardware, or tasks, ensuring they operate with optimized accuracy and efficiency over time.

4. Federated learning

Federated learning is a technique that enables AI systems to learn from distributed data sources whilst ensuring privacy and security. It allows AI to collaboratively train a shared model without transferring sensitive data, preserving privacy and reducing reliance on centralised storage. For example, delivery robots use federated learning to optimise pathfinding without sending raw data, such as sensor inputs or location, to a central server. Instead, they locally update their models and share improvements, preserving both privacy and security.

These developments indicate a key focus on efficiency, adaptability, and learning – all of which are essential for the continued evolution of robotics in complex, real-world environments. Additionally, these advancements contribute to a future where robots collaborate with humans, leveraging their ability to learn from experience and improve over time.

So, what’s next for AI in Robotics?

Just as AI agents are taking over the digital realm, they are about to flood robotics too. AI agents embedded in robotics will supercharge the autonomy and flexibility of robots, enabling them to communicate with humans and even interpret intentions by analysing gestures and potentially emotional cues. Crucial to human-robot interactions, AI agents may prove highly effective in assisted care, hospitality, and other service industries.

Additionally, as technologies like federated learning and edge computing evolve, robots will share knowledge without compromising privacy or relying on centralised data. This will improve scalability and efficiency by reducing the need for costly centralised storage and processing, and enable additional robots to integrate rapidly into existing networks.

So, where does this leave us?

Although there are abundant market opportunities for AI in robotics, the pace at which different markets adopt robotics will vary; with AI being a key factor driving this adoption. Crucial for overcoming challenges related to autonomy, adaptability, and decision-making, AI will empower robots to perform tasks once considered too complex or risky for automation. As AI continues to evolve, it will not only raise important concerns about safety, ethics, and integration but help address them; ensuring robots can work seamlessly alongside humans and contribute to a more productive future.

About the Author:

Holding a BA in Marketing and an MSc in Business Management, Eleanor Wright has over eleven years of experience working in the surveillance sector across multiple business roles.

Der Beitrag A New Era of Intelligent Robots – AI and Robotics erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

SwissCognitive Guest Blogger: Artem Pochechuev, Head of Data and AI at Sigli – “AI for Disabilities: Quick Overview, Challenges, and the Road Ahead”

AI has enormous power in improving accessibility and inclusivity for people with disabilities. This power lies in the potential of this technology to bridge gaps that traditional solutions could not address. As we have demonstrated in the series of articles devoted to AI for disabilities, AI-powered products can really change a lot for people with various impairments. Such solutions can allow users to live more independently and get access to things and activities that used to be unavailable to them before. Meanwhile, the integration of AI into public infrastructure, education, and employment holds the promise of creating a more equitable society. These are the reasons that can show us the importance of projects building solutions of this type.

Yes, these projects exist today. And some of them have already made significant progress in achieving their goals. Nevertheless, there are important issues that should be addressed in order to make such projects and their solutions more efficient and let them bring real value to their target audiences. One of them is related to the fact that such solutions are often built by tech experts who have practically no understanding of the actual needs of people with disabilities.

According to the survey conducted in 2023, only 7% of assistive technology users believe that their community is adequately represented in the development of AI products. At the same time, 87% of respondents who are end users of such solutions express their readiness to share their feedback with developers. These are quite important figures to bear in mind for everyone who is engaged in the creation of AI-powered products for disabilities.

In this article, we’d like to talk about the types of products that already exist today, as well as potential barriers and trends in the development of this industry.

Different types of AI solutions for disabilities

In the series of articles devoted to AI for disabilities, we have touched on types of products for people with different states, including visual, hearing, mobility impairments, and mental diseases. Now, let us group these solutions by their purpose.

Communication tools

AI can significantly enhance the communication process for people with speech and hearing impairments.

Speech-to-text and text-to-speech apps enable individuals to communicate by converting spoken words into text or vice versa.

Sign language interpreters powered by AI can translate gestures into spoken or written language. It means that real-time translation from sign to verbal languages can facilitate communication, bridging the gap between people with disabilities and the rest of society.

Moreover, it’s worth mentioning AI-powered hearing aids with noise cancellation. They can improve clarity by filtering out background sounds, enhancing the hearing experience in noisy environments.

Advanced hearing aids may also have sound amplification functionality. If somebody is speaking too quietly, such AI-powered devices can amplify the sound in real time.

Mobility and navigation

AI-driven prosthetics and exoskeletons can enable individuals with mobility impairments to regain movement. Sensors and AI algorithms can adapt to users’ physical needs in real time for more natural, efficient motion. For example, when a person is going to climb the stairs, AI will “know” it and adjust the movement of prosthetics to this activity.

Autonomous wheelchairs often use AI for navigation. They can detect obstacles and take preventive measures. This way users will be able to navigate more independently and safely.

The question of navigation is a pressing one not only with people with limited mobility but also for individuals with visual impairments. AI-powered wearable devices for these users rely on real-time environmental scanning to provide navigation assistance through audio or vibration signals.

Education and workplace accessibility

Some decades ago people with disabilities were fully isolated from society. They didn’t have the possibility to learn together with others, while the range of jobs that could be performed by them was too limited. Let’s be honest, in some regions, the situation is still the same. However, these days we can observe significant progress in this sphere in many countries, which is a very positive trend.

Among the main changes that have made education available to everyone, we should mention the introduction of distance learning and the development of adaptive platforms.

A lot of platforms for remote learning are equipped with real-time captioning and AI virtual assistants. It means that students with disabilities have equal access to online education.

Adaptive learning platforms rely on AI to customize educational experiences to the individual needs of every learner. For students with disabilities, such platforms can offer features like text-to-speech, visual aids, or additional explanations and tasks for memorizing.

In the workplace, AI tools also support inclusion by offering accessibility features. Speech recognition, task automation, and personalized work environments empower employees with disabilities to perform their job responsibilities together with all other co-workers.

Thanks to AI and advanced tools for remote work, the labor market is gradually becoming more accessible for everyone.

Home automation and daily assistance

Independent living is one of the main goals for people with disabilities. And AI can help them reach it.

Smart home technologies with voice or gesture control allow users with physical disabilities to interact with lights, appliances, or thermostats. Systems like Alexa, Google Assistant, and Siri can be integrated with smart devices to enable hands-free operation.

Another type of AI-driven solutions that can be helpful for daily tasks is personal care robots. They can assist with fetching items, preparing meals, or monitoring health metrics. As a rule, they are equipped with sensors and machine learning. This allows them to adapt to individual routines and needs and offer personalized support to their users.

Existing barriers

It would be wrong to say that the development of AI for disabilities is a fully flawless process. As well as any innovation, this technology faces some challenges and barriers that may prevent its implementation and wide adoption. These difficulties are significant but not insurmountable. And with the right multifaceted approach, they can be efficiently addressed.

Lack of universal design principles

One major challenge is the absence of universal design principles in the development of AI tools. Many solutions are built with a narrow scope. As a result, they fail to account for the diverse needs that people with disabilities may have.

For example, tools designed for users with visual impairments may not consider compatibility with existing assistive technologies like screen readers, or they may lack support for colorblind users.

One of the best ways to eliminate this barrier is to engage end users in the design process. Their opinion and real-life experiences are invaluable for such projects.

Limited training datasets for specific AI models

High-quality, comprehensive databases are the cornerstone for efficient AI models. It’s senseless to use fragmented and irrelevant data and hope that your AI system will demonstrate excellent results (“Garbage in, Garbage out” principle in action). AI models require robust datasets to function as they are supposed to.

However, datasets for specific needs, like regional sign language dialects, rare disabilities, or multi-disability use cases are either limited or nonexistent. This results in AI solutions that are less effective or even unusable for significant groups of the disability community.

Is it possible to address this challenge? Certainly! However, it will require time and resources to collect and prepare such data for model training.

High cost of AI projects and limited funding

The development and implementation of AI solutions are usually pretty costly initiatives. Without external support from governments, corporate and individual investors, many projects can’t survive.

This issue is particularly significant for those projects that target niche or less commercially viable applications. This financial barrier discourages innovation and limits the scalability of existing solutions.

Lack of awareness and resistance to adopt new tools

A great number of potential users are either unaware of the capabilities of AI or hesitant to adopt new tools. Due to the lack of relevant information, people have a lot of concerns about the complexity, privacy, or usability of assistant technologies. Some tools may stay just underrated or misunderstood.

Adequate outreach and training programs can help to solve such problems and motivate potential users to learn more about tools that can change their lives for the better.

Regulatory and ethical gaps

The AI industry is one of the youngest and least regulated in the world. The regulatory framework for ensuring accessibility in AI solutions remains underdeveloped. Some aspects of using and implementing AI stay unclear and it is too early to speak about any widely accepted standards that can guide these processes.

Due to any precise guidelines, developers may overlook critical accessibility features. Ethical concerns, such as data privacy and bias in AI models also complicate the adoption and trustworthiness of these technologies.

Such issues slow down the development processes now. But they seem to be just a matter of time.

Future prospects of AI for disabilities: In which direction is the industry heading?

Though the AI for disabilities industry has already made significant progress in its development, there is still a long way ahead. It’s impossible to make any accurate predictions about its future look. However, we can make assumptions based on its current state and needs.

Advances in AI

It is quite logical to expect that the development of AI technologies and tools will continue, which will allow us to leverage new capabilities and features of new solutions. The progress in natural language processing (NLP) and multimodal systems will improve the accessibility of various tools for people with disabilities.

Such systems will better understand human language and respond to diverse inputs like text, voice, and images.

Enhanced real-time adaptability will also enable AI to tailor its responses based on current user behavior and needs. This will ensure more fluid and responsive interactions, which will enhance user experience and autonomy in daily activities for people with disabilities.

Partnerships

Partnerships between tech companies, healthcare providers, authorities, and the disability community are essential for creating AI solutions that meet the real needs of individuals with disabilities. These collaborations will allow for the sharing of expertise and resources that help to create more effective technologies.

By working together, they will ensure that AI tools are not only innovative but also practical and accessible. We can expect that the focus will be on real-world impact and user-centric design.

New solutions

It’s highly likely that in the future the market will see a lot of new solutions that now may seem to be too unrealistic. Nevertheless, even the boldest ideas can come to life with the right technologies.

One of the most promising use cases for AI is its application in neurotechnology for seamless human-computer interaction.

A brain-computer interface (BCI) can enable direct communication between the human brain and external devices by interpreting neural signals related to unspoken speech. It can successfully decode brain activity and convert it into commands for controlling software or hardware.

Such BCIs have a huge potential to assist individuals with speech impairments and paralyzed people.

Wrapping up

As you can see, AI is not only about business efficiency or productivity. It can be also about helping people with different needs to live better lives and change their realities.

Of course, the development and implementation of AI solutions for disabilities are associated with a row of challenges that can be addressed only through close cooperation between tech companies, governments, medical institutions, and potential end users.

Nevertheless, all efforts are likely to pay off.

By overcoming existing barriers and embracing innovation, AI can pave the way for a more accessible and equitable future for all. And those entities and market players who can contribute to the common success in this sphere should definitely do this.

About the Author:

In his current position, Artem Pochechuev leads a team of talented engineers. Oversees the development and implementation of data-driven solutions for Sigli’s customers. He is passionate about using the latest technologies and techniques in data science to deliver innovative solutions that drive business value. Outside of work, Artem enjoys cooking, ice-skating, playing piano, and spending time with his family.

Der Beitrag AI for Disabilities: Quick Overview, Challenges, and the Road Ahead erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Building AI’s Future – SwissCognitive AI Investment Radar

As major players double down on AI, Coatue Management is leading the charge with a bold $1 billion fund to fuel cutting-edge advancements, while Amazon’s CEO, Andy Jassy, reassures investors that the company’s surging 81% capital expenditures on generative AI will yield returns. Across the globe, China’s DeepRoute.ai secures $100 million to accelerate smart driving technology adoption, signaling rapid growth in autonomous driving.

The Middle East and North Africa (MENA) region also see substantial commitments, with Google dedicating $15 million to boost AI skills, research, and infrastructure through its new AI Opportunity Initiative. Meanwhile, Nvidia and TSMC’s market values surged in October, reflecting the persistent demand for AI chips and the pivotal role of hardware in advancing AI capacities.

As Microsoft pledges an additional $10 billion toward CoreWeave’s GPU infrastructure and Elon Musk courts Middle Eastern investors to raise xAI’s valuation to $45 billion, the global AI landscape is abuzz with momentum. Yet, the wave of Big Tech investments is also testing investor patience as Meta, Amazon, and Microsoft’s hefty AI expenditures begin to weigh on short-term profitability.

Join and explore the critical investments and strategic shifts that define AI’s trajectory, from transformative tech developments to the challenges and promises that lie ahead for investors and innovators alike.

Der Beitrag Building AI’s Future – SwissCognitive AI Investment Radar erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: news.mit.edu – “A Faster, Better Way to Train General-Purpose Robots”

In the classic cartoon “The Jetsons,” Rosie the robotic maid seamlessly switches from vacuuming the house to cooking dinner to taking out the trash. But in real life, training a general-purpose robot remains a major challenge.

Typically, engineers collect data that are specific to a certain robot and task, which they use to train the robot in a controlled environment. However, gathering these data is costly and time-consuming, and the robot will likely struggle to adapt to environments or tasks it hasn’t seen before.

To train better general-purpose robots, MIT researchers developed a versatile technique that combines a huge amount of heterogeneous data from many of sources into one system that can teach any robot a wide range of tasks.

Their method involves aligning data from varied domains, like simulations and real robots, and multiple modalities, including vision sensors and robotic arm position encoders, into a shared “language” that a generative AI model can process.

By combining such an enormous amount of data, this approach can be used to train a robot to perform a variety of tasks without the need to start training it from scratch each time.

This method could be faster and less expensive than traditional techniques because it requires far fewer task-specific data. In addition, it outperformed training from scratch by more than 20 percent in simulation and real-world experiments.

“In robotics, people often claim that we don’t have enough training data. But in my view, another big problem is that the data come from so many different domains, modalities, and robot hardware. Our work shows how you’d be able to train a robot with all of them put together,” says Lirui Wang, an electrical engineering and computer science (EECS) graduate student and lead author of a paper on this technique.[…]

Read more: www.news.mit.edu

Der Beitrag A Faster, Better Way to Train General-Purpose Robots erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: forbes.com  – “The Next Breakthrough In Artificial Intelligence: How Quantum AI Will Reshape Our World”

In the ever-evolving landscape of technology, a new frontier is emerging that promises to reshape our world in ways we can scarcely imagine. This frontier is Quantum AI, the powerful fusion of quantum computing and artificial intelligence. It’s a field that’s generating immense excitement and speculation across industries, from finance to healthcare, and it’s not hard to see why. Quantum AI has the potential to solve complex problems at speeds that would make even our most advanced classical computers look like abacuses in comparison.

Demystifying Quantum AI: The Power Of Qubits And AI

But what exactly is Quantum AI, and why should you care? At its core, it leverages the principles of quantum mechanics to process information in ways that classical computers simply can’t. While traditional computers use bits that can be either 0 or 1, quantum computers use quantum bits or qubits, which can exist in multiple states simultaneously thanks to a phenomenon called superposition. This allows quantum computers to perform certain calculations exponentially faster than classical computers.

Now, imagine combining this mind-boggling computational power with the pattern recognition and learning capabilities of artificial intelligence. That’s Quantum AI in a nutshell. It’s like giving a genius a superpower – the ability to analyze vast amounts of data, recognize complex patterns, and make predictions with a level of accuracy and speed that was previously thought impossible.

What’s particularly exciting is that this technology is becoming increasingly accessible. Tech giants like Microsoft, Amazon, Google, and IBM are now offering Quantum computing as a service. This means that businesses and researchers can tap into the power of quantum computing without having to build and maintain their own quantum hardware.[…]

Read more: www.forbes.com

Der Beitrag The Next Breakthrough In Artificial Intelligence: How Quantum AI Will Reshape Our World erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

SwissCognitive Guest Blogger: Artem Pochechuev, Head of Data and AI at Sigli – “Future Of AI-Powered Solutions For Disabilities: On The Verge Of Fantasy”

Today, while discussing Artificial Intelligence, we often talk about Generative AI tools, virtual assistants, or recommendations assistance. Such tools are already widely adopted and that’s why it is not surprising that they come to our mind in the first turn.

However, the potential of AI is much higher than that. This technology can be used in some mind-blowing solutions that seem to be absolutely fantastic.

Nevertheless, their introduction can be much closer than we may think.

In this article, we offer you to take a look at the most cutting-edge AI-powered projects that can greatly change the lives of people with disabilities (and not only).

Multimodal LLMs

Let’s start with something that sounds the most realistic – multimodal LLMs. Probably, all of you are already well-familiar with models that can work only with text inputs and provide text outputs.

Multimodal models are able to work with data in different formats. It means that they can deal with text, images, and sounds simultaneously and provide a relevant output. That is exactly what GPT-4o is expected to offer.

Of course, such models can be highly helpful for everyone. But their value will be significantly higher for people with different kinds of disabilities, including those with vision impairment, physiological disorders, and mental diseases.

Multimodal LMS can act as full-scale virtual assistants. Their functionality can offer much more possibilities in comparison to well-known solutions like Siri.

What can multimodal LLMs offer to people who can’t interact with their surroundings in a traditional way? We can say “practically everything” and from some point of view, we even won’t exaggerate.

For example, they will be able to explain everything that is written on the screen or describe what is shown in the picture. Their functionality will allow them to instantly translate and read aloud a text from the PDF file. They will help people to interact with their computers and smartphones. Based on the voice command made by users, they will open different menus, choose the necessary options, or move a pointer to the required line, while for a person with low vision or hand tremors, it can be very challenging to do this.

In the future, such models are expected to process video content as well. This will allow them to recognize films and describe their plots for users. Or they will be able to understand what sports game you will show to them and explain the rules.

Of course, these are just a couple of examples that demonstrate how multimodal LLMs can be used by people with disabilities. The range of their applications can be really wide.

AI-powered prosthetics

For people who were born without some parts of their bodies or who lost them under different circumstances, prostheses can become the best solution. These artificial body parts can restore some of the function and appearance of the lost anatomy. However, everything is not as seamless as we may think. The use of traditional prostheses can be associated with huge discomfort and various limitations, like limitations in dexterity or sensory feedback.

Nevertheless, such issues can be at least partially addressed by AI-powered prosthetics. Yes, AI arms today are not just something from a science fiction book. That’s a reality.

Artificial intelligence can significantly enhance the functionality, adaptability, and user experience of prostheses. In such solutions, ML is applied to teach bionic limbs how to understand movement patterns and how to make predictions based on the behaviours demonstrated previously. Thanks to this, limbs become more dexterous and more “natural”.

Such prostheses, both arms and legs, are non-invasive. But they have sensors that can measure electrical signals to identify the user’s intended movement.

Photo: University of Michigan

Of course, the use of AI-powered limbs is much more convenient in comparison to traditional prostheses. AI can automatically adjust artificial limbs for a better fit and can even make real-time changes based on user movements and activity levels.

The most advanced models can provide feedback on pressure and texture, which allows them to simulate the sense of touch for users.

Nevertheless, the cost of such devices is very high at the moment. This is one of the main factors that prevent them from being widely adopted today.

Bionic eye

Bionic limbs are a cutting-edge technology but what do you think about bionic eyes?

These experimental devices can restore functional vision for people who have partial or even total blindness.

The implantation of the earliest version of the bionic eye took place in 2012. The patient who got this artificial eye suffered from profound vision loss. After the surgery, he was able to see light. However, he couldn’t make distinctions within the environment. Since then, this first eye model has been greatly improved. Some other versions helped people start seeing abstract images. Nevertheless, none of the patients has regained vision.

One of the most widely discussed projects from this category is the Prima system by Pixium Vision. Their bionic vision solutions are being developed to help patients with profound vision loss and improve their independence and mobility.

Photo: https://www.pixium-vision.com

The core of their idea is the use of a 2-by-2-millimeter square implant that should be surgically placed under the retina. This implant should receive infrared data from camera-equipped glasses and further turn it into pulses of electricity which will replace signals generated by photoreceptor rods and cones.

Some early feasibility studies conducted in the US and European Union demonstrated that this system could be potentially effective and safe for people. Nevertheless, the project faced some financial difficulties which resulted in the delay in further research and development.

Rehabilitation robots and exoskeleton

Rehabilitation is a very important process for people with disabilities and patients after injuries. AI-powered robots can greatly help in the process of physical therapy through repetitive and controlled movements. They can offer personalized exercises and continuously monitor the progress to optimize recovery outcomes.

Such robots are often used in targeted therapy for patients with neurological or musculoskeletal impairments, such as stroke, spinal cord injury, or orthopedic injuries.

One of the most well-known robots of this kind is Lokomat which helps individuals relearn walking patterns. It ensures the most physiological movement which can be guaranteed by the individually adjustable patient interface.

Photo: https://www.hocoma.com/us/solutions/lokomat

Another type of solution used in rehabilitation is the exoskeleton. Exoskeletons can be defined as wearable devices that work in conjunction with the user’s movements to enhance or assist physical capabilities.

They can help individuals with mobility impairments to stand, walk, or perform other movements. Moreover, they can be used to enhance the physical abilities of healthy individuals, such as in industrial or military applications.

Over the last several years, we could observe the growing interest in designing innovative tools of this kind that incorporate AI. The obvious benefits of such exoskeletons are their capabilities to analyze data and adjust to the individual user’s needs in real-time.

One such groundbreaking AI-powered exoskeletons was developed by a group of researchers at North Carolina State University and the University of North Carolina at Chapel Hill. This wearable device can ensure great energy savings during human movement, which could lead to great improvements in athletic performance and significantly help individuals with mobility issues.

This exoskeleton is powered by data-driven and physics-informed reinforcement learning. With this approach, wearable robots can become intuitive and predict user’s movements.  This technology can also generate synergistic assistance across different activities, such as walking or stair-climbing. The controller can automatically adapt to various kinematic patterns. It means that the transition between activities can take place without any handcrafted control.

Photo: https://www.foxnews.com/tech/ai-driven-exoskeleton-lightens-your-load-elevates-performance

Though the actual prices of exoskeletons can vary from $50,000 to $120,000, Hao Su, Ph.D., associate professor at North Carolina State University and the University of North Carolina at Chapel Hill, noted that their efficient learning-in-simulation framework allows for rapid design and testing in computer simulations.

This can help to reduce the cost of research and development.

“Looking forward, we plan to make our robots truly affordable and accessible through innovative hardware design, namely low-ratio gears and cost-effective but high-torque electric motors. In about one year, we aim to make our exoskeletons for sale at a price range of $1,500 to $4,000, depending on specific features and manufacturing scale,” he explained.

Elderly care robots/ assistive robots for people with disabilities

While talking about robots, we can’t but mention robots that could fully or at least partially replace nurses, tutors, and caregivers.

In August 2023, the first commercial general-purpose humanoid robot Apollo by Apptronik was presented to the public. At the initial stages of its development, it was planned that it would be used in the manufacturing and warehousing industries. Nevertheless, later the range of its use cases was expanded. It can be also helpful in construction, retail, and elderly care. In the latest case, such robots can handle dozens of household chores and become good companions for people who spend a lot of time in isolation due to their disease or disabilities.

Photo: https://apptronik.com/apollo

However, this project is far not the only one in this group.

For example, there are a lot of reports from Japan revealing that the country is actively investing in the automation of elder care by employing various robots.

Though probably the majority of us start thinking about humanoid devices when asked about care robots, it’s far from being true. They can come in different sizes and shapes. Some of them are intended for physical care. In this case, we are talking about those ones that can help lift patients who can’t get on their own. There are robots that assist people in exercising and moving. Some others can track the physical activity of patients, detect falls, and help them use the toilet or take a bath.

There are also robots that are intended to communicate with elderly people, they can entertain them and even conduct cognitive training.

Robot guide dogs

Guide dogs are known to have been helping people with visual impairments for centuries. They can be great assistants and companions but the use of their services is associated with a row of challenges. The training of a guide dog usually requires significant time and expense. Moreover, let’s not forget about an average dog’s lifespan. This explains why a lot of countries face a shortage of trained guide dogs.

For example, according to the data published by the China Association for the Blind, currently, there are only around 400 trained dogs in the country, while the number of people who may need their help is over 17 million.

Dogs require particular care. They all have their personalities. This also can cause some difficulties for people.

But with the application of modern technologies, such issues can be successfully solved. Especially for those who do not feel like having a furry friend, as they can have a robot friend instead.

Robot guide dogs can efficiently provide real-time navigation services for people with visual disabilities and let them travel independently and safely. Such robots can identify road conditions, obstacles, and surrounding facilities. Moreover, they can work with voice prompts and provide vibration feedback, which makes interaction with them quite simple.

It is known that a group of Chinese researchers have been already conducting field tests of a six-legged guide dog that relies on cameras and sensors for navigation. This robot can successfully recognize traffic light signals, while in the case of real dogs, this “feature” is not available.

Photo: https://edition.cnn.com/2024/07/08/china/chinese-robot-guide-dog-intl-hnk/index.html

Of course, robot dogs require some maintenance but at least users do not need to feed them on a daily basis.

Brain-computer interface

Another technology that we should mention is a brain-computer interface. It can establish a direct communication pathway between the brain and an external device. It is possible thanks to its capability to decode the neural signals associated with attempted but unarticulated speech. In other words, it can translate neuronal information into commands capable of controlling external software or hardware systems.

In a very simplified way, we can explain its work as follows:

1. Collection of brain signals using electrodes or sensors;
2. Signal processing, filtering, and amplifying;
3. Extraction of relevant patterns or features within the signals;
4. Translation of these patterns into commands that can be understood by external devices.

Some BCIs are being developed for entertainment purposes. With their help, players can enjoy more immersive experiences. However, the majority of such projects have healthcare-related goals. For example, they can be used to assist in the in the recovery of motor functions.

In this context, it’s worth recollecting Neuralink. That’s definitely one of the most widely-known projects of this kind. This BCI is fully implantable. It’s invisible. And it can help users to seamlessly control their smartphones and computers. This technology can greatly help people with disabilities who are looking for ways to become more independent. Its efficiency in this aspect has been already proven in the first human trial.

Photo: https://neuralink.com/blog/prime-study-progress-update

In January 2024, Noland Arbaugh, a 30-year-old man paralyzed from the neck down, became the first patient who received the Neuralink device. Though there were some technical challenges during the trial, the general results look quite promising.

Thanks to the Neuralink device, the young man got practically full control of a computer. With the power of his mind, he can play games and browse the web at any moment. Moreover, according to Neuralink, Noland has managed to set the human record for cursor control with a brain-computer interface.

In an interview with journalists, Noland explained that the biggest advantage of using a BCI is the possibility of being independent.

“It’s just made me more independent, and that helps not only me but everyone around me. It makes me feel less helpless and like less of a burden. I love the fact that the people around me don’t have to wait for me so much. Outside of being completely healed, I believe what most quadriplegics want is independence,” he said.

Conclusion

Though today the majority of solutions mentioned in this article haven’t been widely adopted, that’s obvious that they have great potential given their incredible social value.

Moreover, we can say for sure that the real power of technologies, and AI in particular, hasn’t been even fully explored yet.

We still have a lot of things to learn and to do. But one thing is clear: today we are close to the future as never before. And we definitely shouldn’t stop in making life easier and better for everyone with the power of AI.

About the Author:

In his current position, Artem Pochechuev leads a team of talented engineers. Oversees the development and implementation of data-driven solutions for Sigli’s customers. He is passionate about using the latest technologies and techniques in data science to deliver innovative solutions that drive business value. Outside of work, Artem enjoys cooking, ice-skating, playing piano, and spending time with his family.

Der Beitrag Future Of AI-Powered Solutions For Disabilities: On The Verge Of Fantasy erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: cio.com – “Adapting to AI regulations in the U.S. and Europe: Impacts on CIOs and global enterprises”

Governments around the world are developing AI laws and regulations, which will bring increasing complexity for CIOs whose businesses are engaged across multiple jurisdictions.

Artificial intelligence (AI) is an increasingly large portion of information technology investments and societal discussions. Many governments have started to define laws and regulations to govern how AI impacts citizens with a focus on safety and privacy; IDC predicts that by 2028 60% of governments worldwide will adopt a risk management approach in framing their AI and generative AI policies (IDC FutureScape: Worldwide National Government 2024 Predictions). This article focuses on nascent regulations in Europe and the U.S. and implications for CIOs.

AI regulations in Europe

In late 2023 the European Union (EU) created a draft AI Act, which was subsequently approved by the EU Parliament on March 13, 2024. As one member noted, the EU now has the first binding law on artificial intelligence that will protect the human rights of workers and citizens. The regulation will fully come into effect 24 months after its publication. The Act balances the need to protect democratic rights, rule of law, and environmental sustainability while encouraging innovation, particularly in Europe. AI applications that threaten citizens’ rights, such as predictive policing or untargeted scraping of internet facial images, are banned. Similarly, law enforcement’s use of biometric information systems is prohibited.

The EU AI Act will require member states to create a database of high-risk AI systems to monitor activities in the EU market. National governments will be required to enforce regulations and monitor AI market developments.[…]

Read more: www.cio.com

Der Beitrag Adapting to AI regulations in the U.S. and Europe: Impacts on CIOs and global enterprises erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: medium.com – “Wisdom From The Women Leading The AI Industry, With Dalith Steiger-Gablinger of SwissCognitive”

Understanding the Power of Technology: It’s crucial for women in AI to grasp the potential of technology and articulate its implications for their businesses. This understanding enables effective communication and decision-making in this tech-driven world.

As part of our series about the women leading the Artificial Intelligence industry, we had the pleasure of interviewing Dalith Steiger-Gablinger.

Dalith Steiger-Gablinger is an AI investor, serial entrepreneur, and co-founder of SwissCognitive. As an AI strategist, innovator, and investor, she advises various companies and actively participates in boards, AI accelerator programs, and mentors young girls in tech. Dalith delivers insightful keynotes worldwide, drawing from her extensive experience and knowledge. Beyond her professional endeavors, she finds joy in mountain biking and spending time with family and friends.

Thank you so much for joining us in this interview series! Can you share with us the ‘backstory” of how you decided to pursue this career path in AI?

Thank you for having me! I studied mathematics and information technology, although I faced many challenges with math during my schooling, leading me to drop out temporarily due to a lack of understanding. Upon returning and taking a mathematics course with a different teacher, I was able to grasp the subject. This experience taught me the importance that clear communication is key to understanding, and if someone doesn’t understand something, it’s the mistake of the messenger, not the receiver. This became an important backbone to my future career pursuits in AI.

I then started a career in the banking industry as a software developer 28 years ago. I was just one of the two women in this male-dominated field in the Zurich area, where I quickly became valued for my ability to bridge communication gaps between the IT and business teams.

About 10 years ago, I then attended a Gartner event in Barcelona where I was introduced to the concept of the “Digital Employee”. The Digital Employee is a conversational system powered by AI that can interact with users in order to accomplish a wide range of business tasks. Recognizing its immense potential, I was inspired to start my journey into the complex, and growing AI space.

And it turns out I wasn’t the only Swiss professional becoming drawn into the field of AI. The unique talents and attributes of Switzerland’s business community meant that we had the brain and technology power to quickly become globally competitive. This realization prompted the founding of SwissCognitive, where I currently work as a Global AI Strategist, facilitating knowledge exchange among companies and advise them on AI strategies and investments.

What lessons can others learn from your story?

The pivotal lesson from my experience is to challenge assumptions about oneself and the information received. Many women may doubt their understanding, but it’s essential to challenge this mindset and advocate for clearer communication, especially in the world of emerging technologies. Remember, if something isn’t clear, it’s not solely the individual’s fault; instead, it’s crucial to prompt the communicator to explain in simpler terms.

One instance where I shifted from self-doubt to challenging the communicator was when I joined a select group of AI fund investors as an expert to help explain the AI landscape. Despite having ongoing meetings with older male investors and initially refraining from expressing my opinions, I eventually felt compelled to speak up in a particular situation. Although I felt insecure at that moment, when I did speak up, the investors acknowledged the value of my insights and responded by saying “Wow, you’re so right.” And then the dialogue flowed much more effectively and productively from there!

It’s important to keep in mind that encouraging dialogue with phrases like “Am I right?” fosters an environment where everyone can contribute and question the message. The transformation in Switzerland’s history of women’s voting is very recent compared to other counties, so the female voice is still growing here! In my role, now supporting women and entrepreneurship in this critical AI space, I am always encouraging women of the need to challenge assumptions and promote equality.

Can you tell our readers about the most interesting projects you are working on now?

Currently, one of the most engaging projects I’m involved in is my ongoing effort to empower women entrepreneurs and the younger generation by raising awareness about the exciting potential of AI. Another one would be supporting AI startups in their strategy, growth and market entry. These activities include participating in keynote panels, engaging with students at universities, and connecting with the vibrant community we lead of over half a million AI enthusiasts on platforms like LinkedIn.

It’s crucial to invest in the younger generation and recognize the responsibility of the older generation in shaping their future. The intersection of that idea with cutting-edge technologies like AI is what I find most interesting today. I do this work through educating, giving lectures, and creating forums for collaboration and ideation.

Another dear project to me, is unleashing investments for AI startups. We mix and match these startups with investors, VC funds and corporate ventures to boost their success. Especially in Switzerland, we are proud to see that more and more Swiss investors are willing to invest into this particular growing tech scene. We see Switzerland turning more and more from a heritage culture to an investment culture, believing in its startup scene and enabling our innovation nation to become a startup nation.[…]

Read more: www.medium.com

Der Beitrag Wisdom From The Women Leading The AI Industry, With Dalith Steiger-Gablinger of SwissCognitive erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.