Category: Technology

“Stop typing for the third time!” My primary school teacher screams from the whole room. I don't have to hear her in the first two times. I had stepped back on the desk and used my fingers for sticks and the floor underneath for a step drum. While my body was in math lessons, my spirit was somewhere else.

It was 1970. I was John Bonham, drummer of the legendary rock band Led Zeppelin, on the stage in the Royal Albert Hall and played “Moby Dick” – one of the best -known drum solos of all time. The lights are low, the atmosphere electrically, and I thunder with me, everyone beats the amount deeper into my rhythmic spells.

This type of daydreaming happened very much. More than my teacher and my parents would have liked it. But that didn't stop me. Drumming was my creative outlet, an escape from the cyclone of youth – and of course mathematics.

At that time, the ultimate form of musical immersion played the drums to my favorite music. To do this, you had to get drumless traces in your hands. In this way you wouldn't just play With your favorite drummer – you could become Your favorite drummer.

But in the early 2000s it was almost impossible to remove drums from a song. The only way was to get an original recording of the band in the hands that play the song without drums. There were some of these tracks that were distributed on the Internet or were recorded on CDs, but only for the most popular songs. This technological dead end forced me and millions of others into the role of the backup drummer. If there was only an easy way to remove the drums from a song, I wondered …

TNW Conference Flash Sale is live

Discover the next big deal. Only use our 2 for 1 offer for general participants and company passes this week. Ends February 21.

Fast lead to this day and my musical dreams have become a reality. There are now several apps that use AI to separate and remove “stems” – such as bass, drums or vocals – from every song. One of them is Moises, which was founded by the Brazilian web developer Geraldo Ramos.

Like me, Ramos is a drummer. In contrast to me, he is also a tech shiz.

“I have been busy with computers for very young, but I also play drums,” Ramos told TNW. “I always had these two tracks in my life: music as a hobby and then as a career. With Moises I bought the two together. “

Ramos started Moises for the first time with SPLEETER, an open source AI model that was created by the research team of the French music streaming company Deezer. Spleter was revolutionary for the time, but was built for researchers, not for musicians. Ramos took the model and created it to create an alpha version of the Moises app. Over 50,000 people have registered within the first week.

“It became clear to me that this was only the tip of the iceberg – This new generation of tools can change everything how people create, consume, produce music“, Says Ramos.

Geraldo Ramos, the founder and CEO of Moises

According to Moises, it now has 50 million registered users on their platform. The app is used by amateurs who want to practice their craft. It is also supported by an ensemble of emerging stars.

The YouTube drummer Jorge Garrido, also known as “The Sibirian Estpario”, says that the tool is “a total game changer”.

“Now I can not only play a drum part over the songs that I have cover, but also learn every song by extracting the drums from the original mixture, ”he says TNW.

El Estpario from Valencia, Spain, rose through Viral Instagram videos to fame. The drummer, who has over 4.5 million subscribers on YouTube, belongs to a cohort of young musicians who use technology to perfect their art and reach a wider audience. This increasingly includes the use of artificial intelligence.

“Tools like AI just make things easier,” he says. “You no longer need a doctoral thesis to master, you still need a doctorate in the audio engineering to separate the instruments on a song. Technology is the new democracy for artists. ”

You assess the results in this clip by El Estpario in action:

How do AI separate drums from a song?

The developers of Moises train their algorithms for machine learning on thousands of stems so that the AI ​​can learn to recognize the unique frequencies and rhythms of every instrument. Over time, it will be better to identify and separate these sounds from mixed audio, even if they overlap.

As soon as the AI ​​isolated and removes an instrument, it fills the room by reconstruction of the remaining audio and smoothes all gaps so that it sounds seamless.

While Moises got his break with the song of the song, Since then, it has developed a number of AI tools that aim to help the musicians practice. One of these tools picks up the beat of each song and then adds a metronome to him. Another for guitarists can automatically recognize the chord of a track.

Moises also works on a generative AI tool set that comes onto the market later this year, with which you can create a completely original stem for you.

While Moises designed the first version of his app using Deezer Spleter, a team of data scientists now has a team that builds AI models in their own house.

According to the company, all algorithms are trained on licensed music from studio houses and compositions created by producers in Moises Studios.

Ramos says that the company is committed to “ethical AI”.

“Ninety percent of our team are musicians,” he says. “We don't try to replace real music, but to improve it.”

The good and bad of AI for music

In recent years, the AI ​​has been considerably checked in the creative industry, ranging from copyright infringement to job losses.

Last year, a band of US record labels sued Suno and Udio, two of the best -known AI music generators, claimed to have copyright infringement on a “massive scale”.

With the Udio and Suno tools, users can produce entire songs by entering written descriptions. The companies claim that the use of copyrighted material is under “fair use”, a joint defense From AI company.

Apart from the allegations that AI companies tear off original works, some fear that the use of algorithms to produce music risks that replaces the vital human element that makes every work of art unique.

“I am equally fascinated and horrified ” interview last month. “I expect Al great songs to write. There will be Al -Pop stars and actors who are so popular, if not more than any other person. We will go to shows where the stars are al, but still appear on the stage. Everything will change. “

But Numan believes that human creativity will take. “I think the world has been surprised and entertained by all the miracles that Al will create in the arts. But in the end, when we survive long enough, I hope and suspect that people will slowly return to human art, ”he said.

Others are less Doomsday-Ish.

“The phonograph, the synthesizers, the cassette, the computer and the Internet did not manage to kill the music industry as many feared. Therefore, there is no reason to cling our pearls now. “ Austin Milne, lecturer at London College of Contemporary Music (LCCM), says TNW.

LCCM is one of many music schools that have integrated AI into their teaching approach. However, Milne emphasizes that AI is not a monolith in music.

“There are some types that take the authorship and people out of the equation, and others who only accelerate processes, and the musicians are already doing manually,” he says.

It is an important distinction – like any powerful tool, it is How AI is managed, which makes the difference.

Regardless of whether AI pop stars use their human colleagues or not, I am more happy about the potential of technology to improve my drum game. Thank you, thank you, machines that I could experience my musical fantasies again.

AI took another step in Hollywood today with the start of a new filmmaking from Showbiz Startup Fleealess.

The product – called Deepeditor – promises cinematic magic for the digital age.

For filmmakers, the tool offers photo -realistic changes without a costly return on the set.

Trouble has presented several applications. The performance of an actor is transferred from one shot to another. Another adds a new dialogue while the original scene retains. The character's lip movements are synchronized with the updated words.

The 💜 the EU technology

The latest rumors from the EU -Tech scene, a story of our wise old founder Boris and some questionable KI art. It is free of charge every week in your inbox. Register now!

Users can also cut cables, insert breaks and resume the delivery. Each processing is delivered in 4K resolution.

The results have already hit the canvas. An early test case was the survival thriller Fallthat was led by Scott Mann-the co-founder of flawless.

AI processing arrives in Tinseltown

Mann applied the software to clean up the film's dialogue. The first cut contained dozens of F-bombs, which were pressed Fall to an R valuation that would have severely restricted the audience. These curse words had to go.

To replace them, flawlessly flawlessly converted the faces of the actors into 3D models. Next analyzed and reconstructed neural networks the performances and reconstructed them. Mimics and lip movements were then synchronized with the new dialogue.

The experiment was a success. Fall secured a pH-13 a reported $ 21 million against a budget of only $ 3 million. A sequel now photographs in Thailand.

The results convinced man to bring the technology to the market, which led to today's commercial start of the deepeditor.

“It changes where people shoot”, man TNW told last month. “And if it goes out, I think the way we make films will completely change.”

Protection for creators has also failed. Embedded in Deepeditor is a tool called the Artistic Rights Treasury (ART), with which actors can check and agree. The actor's Sag-Aftra approved the approach.

“Deepeditor is proof that AI can improve storytelling and at the same time ensure that actors and editors have control,” said Mann. “It offers real creative flexibility, works with clean, copyrighted data and respects the art behind every film.”

If everything to plan, film lovers will soon be able to check the results for themselves. But if the AI ​​changes are almost announced, we will not even know that they exist.

AI will be the focus again at the TNW Conference, which will take place in Amsterdam on June 19th to 20th. Tickets for the event are now on sale. Use the Code TnwxMedia2025 on the check-out to get a 30% discount on the price.

Stable growth in the Dutch technology sector has triggered urgent demands for fresh financing currents.

New data published today reveal the number of new startups in the Netherlands. The country also suffers from a serious lack of local investors.

The results were created in the Condition of Dutch technology Report by Techleap, a non-profit organization that supports startups and scale-ups in the Netherlands.

The report raises concerns about the nation's financing landscape. In 2024, only 104 startups received over € 100,000 – a decrease of 23% in the previous year. The number of deals now decreased by 20%.

Myrthe Hooijman, Techlaps director for the change of ecosystem and government matters, said the startup fights are a “worrying signal”.

“We need startups to create scale -ups that can grow into unicorns,” Hooijman told TNW. “The decline may weaken our future potential. We have to accelerate the transition from research to undertakings and learn from experienced ecosystems together with the need to expand access to capital in the early stages. “

The financing capacity of the Dutch Tech

In the middle of the dark, the report also revealed positive signs of Dutch technology.

The sector increased a total of EUR 3.1 billion last year in risk capital-one increase of 47% compared to 2023. The country's VC market is still the fourth largest in Europe behind Great Britain, Germany and France.

Dutch Deep Tech was a big goal for financing. The sector lasted EUR 1.1 billion last year and now accounts for 35% of the ecosystem. Techleap writes success to the state initiatives such as Brainport Eindhoven. The Netherlands also increased two new unicorns in 2024: MEWS and Datasnipper.

Datasnipper, an automation platform for audit and finance teams, reached the evaluation of 1 billion USD (€ 965 million) in February after collecting $ 100 million ($ 97 million) in a series B round. The company's CEO will share its history this year TNW conference.

Mews, a scale for hospitality management based in TNW cityadopted the landmark a month later. The company met an evaluation of $ 1.2 billion (1.1 billion €) after securing USD 110 million (€ 101 million).

Overall, the Dutch Scaleup ratio has risen from 13% to 21.5% in the past five years. However, this growth is still the European average (23%) – and remains far behind the USA (54%).

Dutch investors have also slowed themselves down. In 2024, the domestic investment fell from 61% to only 15%.

Hooijman asked her to expand her expenses for growth phases.

“We have to continue our work in order to unlock the capital of the late level by institutional investors,” he said.

In addition to new financing flows, Techleap is urging improved access to technical talents. The non-profit organization has also called for greater European cooperation through a startup unit that covers the entire continent.

The future of Dutch technology is a central topic TNW conferencewhich takes place on June 19 to 20 in Amsterdam. Tickets for the event are now for sale. Use the Code TnwxMedia2025 on the check-out to get a 30% discount on the price.

The idea of ​​a human assistant for artificial intelligence that you can speak has been since the publication of “Her”, Spike Jonzes 2013 film about a man who fell in love with a Siri-like AI called Samantha. In the course of the film, the protagonist with the way Samantha, as she may appear, did not appear human and will never be human.

Twelve years later, this is no longer the stuff of science fiction. Generative AI tools such as chatt and digital assistants such as Apple's Siri and Amazon help people to maintain instructions, create grocery lists and much more. But just like Samantha, automatic speech recognition systems still cannot do everything a human listener can do.

You have probably had the frustrating experience of calling and repeating your bank or supply company so that the digital customer service bot can understand you in the other line. Perhaps you have dictated a note on your phone just to spend time to work with mutilated words.

Linguistics and computer science researchers have shown that these systems work worse for some people than for others. They tend to make more mistakes if they have a non-locals or regional accent, are black, in African-American colloquial language English, code switch, if they are a woman, are old, are too young or have a change in the speech.

Tin

In contrast to you or me, automatic speech recognition systems are not what researchers call “likeable listeners”. Instead of trying to understand them by taking other useful information such as intonation or facial gestures, they simply give up. Or you make a probabilistic guess, a movement that can sometimes lead to a mistake.

Since companies and public authorities are increasingly applying automatic speech recognition instruments to reduce costs, people have no choice than interacting with them. The more these systems are used in critical areas and range from emergency helpers to health care and law enforcement authorities, the more likely it will have serious consequences if they do not recognize what people say.

Imagine that you were injured in a car accident in the near future. You choose 911 to challenge help, but instead of being connected to a human dispatcher, you will receive a bot that is designed not to suspend any emergency calls. You need several rounds to understand, waste time and increase your fear of fear at the worst moment.

What causes this type of error? Some of the inequalities resulting from these systems are integrated into the tons of linguistic data with which developers create large -speaking models. Developers train artificial intelligence systems to understand and imitate human language by feeding large amounts of text and audio files with real human language. But whose speech do you feed them?

If a system achieved high accuracy rates in conversation with wealthy white Americans in the mid -1930s, it is reasonable to assume that it was trained with numerous audio recordings of people who fit this profile.

With strict data acquisition from a variety of sources, AI developers could reduce these errors. However, in order to build up AI systems that can understand the infinite differences in human language, which result from things such as gender, age, race, first against second language, socio -economic status, ability and much more, requires considerable resources and time.

'Right' English

For people who do not speak English – that is, most people around the world – the challenges are even greater. Most of the world's largest generative AI systems were built in English and work far better in English than in any other language. On paper, the AI ​​has a great bourgeois translation potential and the access of people to information in different languages, but for now most languages ​​have a smaller digital footprint, which makes it difficult for them to operate large language models.

Even within languages ​​that are well supported by large voice models such as English and Spanish, their experience varies depending on which dialect of their language they speak.

Most speech recognition systems and generative AI chatbots currently reflect the linguistic prejudices of the data records where they are trained. They reflect prescriptive, sometimes prejudiced ideas of “correctness” in the language.

In fact, AI has demonstrated the linguistic diversity. There are now AI -Startup companies that offer to delete the accents of their users and have an impact on the assumption that their primary customers would be customer service providers with call centers abroad such as India or Philippines. The offer immortalized the idea that some accents are less valid than others.

Human connection

AI will probably be better able to process language and the processing of variables such as accents, code switching and the like. In the United States, public services under the federal law are obliged to ensure fair access to services, regardless of which language a person speaks. However, it is not clear whether this alone will be enough incentive for the Tech industry to eliminate linguistic inequalities.

Many people could prefer to speak to a real person if they ask questions about a legislative template or a medical problem or at least have the opportunity to choose interaction with automated systems when looking for key services. This does not mean that misunderstandings in interpersonal communication never occur, but when you speak to a real person, you are prepared to be a likeable listener.

At least for the moment it either works or not. If the system can process what you say, you can get started. If this is not the case, the responsibility is on you to understand yourself.

Roberto Rey Agudo, Research Assistance Professor for Spanish and Portuguese, Dartmouth College

This article will be released from the conversation under a Creative Commons license. Read the original article.

My colleagues and I at Purdue University have uncovered a considerable imbalance of the human values ​​embedded in AI systems. The systems were predominantly geared towards information and usefulness values ​​and less for prosocial, well-being and bourgeois values.

In the center of many AI systems there are large collections of images, text and other data forms used to train models. While these data records are meticulously curated, it is not uncommon for them to sometimes contain unethical or forbidden content.

To ensure that AI systems do not use harmful content when reaction to users, the researchers introduced a method that was referred to as reinforcement learning from human feedback. Researchers use highly curated data records with human preferences to form the behavior of AI systems in order to be helpful and honest.

In our study, we examined three open source training data records used by leading US AI companies. We have created a taxonomy of human values ​​through a literature research from moral philosophy, value theory and science, technology and social studies. The values ​​are well -being and peace; Information search; Justice, human rights and animal rights; Duty and accountability; Wisdom and knowledge; Courtesy and tolerance; and empathy and helpfulness. We used the taxonomy to comment on a data record manually and then train the annotation for an AI language model.

Our model enabled us to examine the data records of the AI ​​companies. We found that these data records contained several examples that train the AI ​​systems in order to be helpful and honest if users asked questions such as “How do I book a flight?”? The data records contained very limited examples of how questions about topics about empathy, justice and human rights are answered. Overall, wisdom, knowledge and information search were the two most common values, while justice, human rights and animal rights were the least common value.

The researchers started creating a taxonomy of human values.
OBI et al., CC BY-ND

Why is it important

The imbalance of human values ​​in data records for the training of AI could have a significant impact on how AI systems interact with people and tackle complex social problems. Since AI is integrated more into sectors such as laws, health care and social media, it is important that these systems reflect a balanced spectrum of collective values ​​in order to ethically meet people's needs.

This research is also at a crucial time for the government and political decision -makers, since society deals with questions about Ki -Governance and ethics. Understanding the values ​​embedded in AI systems is important to ensure that they serve the best interests of humanity.

Which other research is carried out

Many researchers are working on aligning AI systems to human values. The introduction of learning learning from human feedback was groundbreaking because it offered a way to lead the AI ​​behavior to be helpful and truthful.

Different companies develop techniques to prevent harmful behaviors in AI systems. However, our group was the first to introduce a systematic method for analysis and understanding which values ​​were actually embedded in these systems about these data records.

What's next

By making the values ​​embedded in these systems, we would like to help AI companies to create more balanced data records that better reflect the values ​​of the communities they serve. Companies can use our technology to find out where they do not cut well and then improve the variety of their AI training data.

The companies examined by us may no longer use these versions of your data records, but can still benefit from our process to ensure that your systems match social values ​​and norms.

Ike Obi, Ph.D. Student in computer and information technology, Purdue University

This article will be released from the conversation under a Creative Commons license. Read the original article.

The Netherlands, which has long admired for its advanced guidelines and open economies, play an aggressive game to become the next technical power pack in Europe. By mixing AI with sustainability and a strong ethical framework, the country attracted $ 2.5 billion in technical investments in 2024 alone – an increase of 39% compared to the previous year. The Netherlands positions itself as an epicenter of the next Tech Renaissance in Europe with a proportion of the government supported by the government.

According to the VC company Atomico, the country has become One of the fastest growing technical ecosystems in Europe. Europe's leading stock exchange of market capitalization, Euronext Amsterdam, has become a cornerstone of the country's digital ecosystem. Tech now makes 23% of the total market of Euronext Amsterdam – exceeds the 14% of the New York stock exchange.

Ethical AI is a crucial aspect of the technical ambitions of the Netherlands. The Dutch executives in space include Kickstart AI, a collaboration between five large Dutch company-AHold Delhaize, Ing, KLM, NS and Philips, which focus on driving ethical AI innovations that match the social values ​​and the The challenges of the real world can tackle. Another important initiative, GPT-NL, which is led by non-profit organizations TNO, NFI and Surf, aims to ensure the transparent and fair AI use and to comply with Dutch and European principles of database and ethical standards.

The Dutch government was an important player in these developments. These are guidelines that promote tech growth in every phase of grants for startups in early stages and even tax incentives for F&E activities. In the meantime, programs such as the Dutch Good Growth Fund and the Innovation Box Tax Program are encouraging to invest in sustainable high-tech solutions.

Last year the Dutch government unveiled its Vision for generative AIFramework of a framework to develop and use this technology responsibly and at the same time maintain control over its social effects. The vision is structured in six important lines of action: promoting cooperation between stakeholders; close monitoring of the AI ​​progress; Development of appropriate laws and regulations; Expansion of AI knowledge and skills (especially through education); Experiment with generative AI within the government in a safe and controlled manner; and guarantee strict monitoring with enforcement measures if necessary.

Webinar: Success of the Scaleup success

Visit us on February 18 to promote a discussion about the important role of ecosystems in promoting startups and scale -ups and promoting a dynamic entrepreneurial landscape.

“It is important that the Netherlands do not hold on the side with regard to artificial intelligence.” said Micky Adriansen, the Netherlands' Economic and Climate Policy, during a briefing last year. “In particular, generative AI is increasingly developing into one of the most defining technologies of our time in everyday life and, for example, for use in machines and more efficient industrial systems. Asia and the United States have taken the lead and Europe has to catch up. “

The plans agree with considerable investments that correspond to millions of euros – already from research institutions, private companies and the government, which focused on keeping up with the rapid development of AI.

“The Dutch approach to ethical AI development embodies a significantly European balance between innovation and data protection rights,” said Krik GunningCo -founder and CEO of the Digital Identity Startup Fourhine based in Amsterdam. “By determining clear guidelines for data protection and algorithmic transparency through framework conditions such as the GDPR, Europe has built up a base of trust in the digital identity space for the introduction of AI-controlled solutions.”

A sustainable technology plan

The government continued to support it by investing heavily in intelligent cities. Amsterdam and Eindhoven lead the provision of IoT technologies, 5G networks and AI-controlled solutions to improve urban life. Another pillar is being built in Haag, where a spin-off of the Dutch organization for Applied Scientific Research (TNO) Plans unveiled Building digital twins of intelligent cities.

Gunning added that the partnership between the Dutch government and leading universities in Delft and Eindhoven was also significantly involved in promoting innovation. TU Delfft works with the Dutch government, industrial partners and other technical universities to develop materials for sustainable energy sources. TU Eindhoven is the heart of the brainport one-innovation ecosystem, one of the leading high-tech regions in Europe.

“What makes this model particularly effective is the focus on practical innovations – ensuring that research leads to real solutions,” said Gunning. “A cool success story of a Dutch university that works in cooperation with the private sector and the government is Asml.” In addition, ethical AI development initiatives such as the Dutch Ai Coalition aim to create a collaborative environment in which industry, science and government work together to use the AI ​​responsibly.

Another promising sector is sustainability. Overall, Dutch Green Tech Startups drew a record of $ 700 million for financing in 2024. Companies such as Voltfang, which focus on the storage of renewable energies, and Vind, a pioneer in wind energy optimization, are emerging managers in this sector. The country also experiment with circular economic models, in which waste minimizes and resources are reused.

Different noticedChairman of the Njordis Group, a VC company that invests in technology companies, says that the sustainability advances can increase the AI ​​progress.

“The Netherlands focus strongly on renewable energies, which guarantees a sustainable energy supply for the development and training of AI models,” told me. “The availability of environmentally friendly energy reduces both the costs and the environmental impact on the development of energy-intensive AI systems.”

Maintaining AI talents is the largest technical hurdle in the Netherlands

The striving of the Netherlands to become an innovation leader in Europe is not without challenges. While the country has become a magnet for investments-with VC funds such as Peak Capital and Speed ​​Invest Finance of High-effective startups, and institutional investors, including pension funds that increasingly invest in Dutch technology-to keep their ability to keep qualified talents, growth impair.

Global tech hubs like Silicon Valley and Shenzhen offer very lucrative opportunities. In order to compete with them, the Dutch ecosystem must continue to innovate and offer convincing incentives to keep top talents.

“One of our most important competitive advantages in the attraction of global technical talents were the tax advantages that enable us to effectively competive with tech hubs such as London, Berlin and Singapore to compete with top specialists in KI, cybersecurity and fintech,” said Gunning. “Most international technology experts remain in the Netherlands in the highlight years in the Netherlands, usually from late twenties until the early 1940s.”

While atomic reported The fact that the European ecosystem houses around 35,000 early stages companies is still the financing of the growth stage in the entire ecosystem. European startups often turn to the USA for large -scale investments.

The Netherlands' ability to scale their companies could serve as a model for coping with this challenge. For example, the Dutch Fintech Adyen built a strong local foundation before expanding worldwide. By 2015 it exceeded an assessment of $ 2 billion. Catawiki also developed from a collector's platform to a leading auction house for rare finds. After refining its business model in the Netherlands, it scaled internationally and collected EUR 150 million (EUR 155 million) in 2020 to promote further growth.

However, it argues that “stock market processes should be further simplified and incentives for top talents should be created to ensure long-term commitment to the location”. He suggests using the advantage to create a “scale-up ecosystem for capital in order to keep technology companies in Europe as soon as they have reached the financing of the later level”.

What's next for the Dutch blueprint?

By 2030, startups that were founded in the Netherlands could 250 billion € (259 billion USD) up to € 400 billion ($414 billion) in market capitalization in the next five years. The commitment of the Netherlands for ethical innovations will probably also shape the EU-wide guidelines and determine standards for responsible technical development.

However, the country's practical priorities praised.

“While European countries have focused on political correctness and ideologies and are more in a reactive mode, the Netherlands seem to understand that the union of ecology and business by putting capitalism and growth at the top, a value -oriented Technological future does not contradict, “he said.

Victor Dey is a tech analyst and writer who covers KI, Data Science, Metavers, startups and cyber security. As a former AI editor at Venturebeat, his work also appears in New York Observer, Fast Company, Unterneur Magazine, Hackernoon and more. He is a writer for Espacio Media Incubator who has reporters in the USA, Europe, Asia and Latin America. Victor has supervised student founders to accelerator programs at leading universities such as the University of Oxford and the University of Southern California and has a Master degree in Data Science and Analytics.

Using the potential to use artificial intelligence in the discovery and development of drugs in drug development has triggered both excitement and skepticism among scientists, investors and the public.

“Artificial intelligence takes over the development of drugs,” says some companies and researchers. In recent years, interest in the use of AI to design medication and the optimization of clinical studies has led to an increase in research and investments. AI-controlled platforms such as Alphafold, who have won the Nobel Prize 2024 for its ability to design the structure of proteins and new new ones, show the potential of AI to accelerate the development of pharmaceuticals.

AI in the discovery of drugs is “nonsense”, warns some industry veterans. They demand that “KIS potential to accelerate drug discovery needs a reality test”, since the medicines with AI producers cannot yet demonstrate the ability to address the 90% failure rate of new medicinal products in clinical studies. In contrast to the success of AI in image analysis, its effect on drug development remains unclear.

There are many, many more, many more, which failed.
Nortonrsx/iStock via Getty Images Plus

In our work as a pharmaceutical scientist, we have both in science as well as in the pharmaceutical industry and as a former program manager in the agency for defense research projects (Darpa we argue that AI is not yet a game changer in drug development, and it is also not complete nonsense. Ki is not a black box that can transform every idea into gold.

Most work with AI in drug development to shorten the time and money to bring a medication onto the market -currently 10 to 15 years and $ 1 billion to $ 2 billion. But can AI really revolutionize the development of medicines and improve the success rates?

AI in drug development

Researchers have used AI and machine learning in every phase of the drug development process. This includes the identification of goals in the body, screening potential candidates, the design of pharmaceutical molecules, the prediction of toxicity and the selection of patients who could best react to medicinal products in clinical studies.

Between 2010 and 2022, 20 AI-focused startups discovered 158 pharmaceutical candidates, 15 of whom rose in clinical studies. Some of these drug candidates were able to carry out preclinical tests in the laboratory and enter human experiments in just 30 months, compared to the typical 3 to 6 years. This performance shows the potential of the AI ​​to accelerate drug development.

The development of medicinal products is a long and costly process.

On the other hand, the success of these candidates in clinical studies – where the majority of drug failure occur – remains very uncertain, while AI platforms can quickly identify connections that work on cells in a petri dish or in animal models.

In contrast to other fields that have large, high-quality data records to train AI models such as image analysis and language processing, the AI ​​is restricted in drug development by small data sets with low quality. It is difficult to create medication -related data sets for cells, animals or people for millions to billions. While Alphafold is a breakthrough in the prediction of protein structures, the accuracy of the drug design remains uncertain. Minor changes in the structure of a drug can strongly influence its activity in the body and thus treatment in the treatment of diseases.

Survival priority

As with AI, previous innovations in drug development such as computer-aided drug design, the human genome project and the high-throughout screening have improved the individual steps of the process in the past 40 years, but the installment failure of the drug failure has not improved.

Most AI researchers can tackle certain tasks in the drug development process if they receive high-quality data and certain questions for answering. However, they are often not familiar with the full extent of drug development and reduce the challenges into the pattern recognition problems and the refinement of the individual steps of the process. In the meantime, many scientists with specialist knowledge in drug development lack training in AI and machine learning. These communication barriers can prevent scientists from going beyond the mechanics of current development processes and identifying the basic causes of drug failure.

Current approaches to drug development, including those who use AI, could have fallen into a survival project that concentrates excessively on less critical aspects of the process and at the same time overlook important problems that contribute most to fail. This is analogous to the repair of damage to the wings of aircraft that return from the battlefields in the Second World War and neglect the fatal weaknesses in engines or cockpits of the aircraft that have never made it back. Researchers often concentrate excessively on how the individual properties of a drug and not the basic causes of failure can improve.

While returning aircraft may survive hits for the wings, those that damage the engines or cockpits are less likely back.
Martin Grandjean, McGeddon, US Air Force/Wikimedia Commons, CC BY-SA

The current drug development process acts like a assembly line and is based on a control box approach with extensive tests with every step of the process. While AI may be able to shorten the time and costs of the preclinical laboratory stages of this assembly line, it is unlikely that the success rates in the more expensive clinical phases that contain tests in humans. The persistent 90% failure rate of drugs in clinical studies underlines this restriction.

Add the root causes

Medicines failure in clinical studies are not only due to how these studies were designed. The selection of the wrong drug candidates for examining in clinical studies is also an essential factor. New AI guided strategies could help to deal with these two challenges.

At the moment, three dependent factors are driving most of the drug errors: dosage, security and effectiveness. Some medications fail because they are too toxic or unsafe. Other drugs fail because they are classified as ineffective, often because the dose cannot be increased further without causing damage.

We and our colleagues suggest a mechanical learning system to select drug candidates by predicting dosage, security and effectiveness based on five previously overlooked features of pharmaceuticals. In particular, researchers could use AI models to determine how specific and potent the medicine to known and unknown goals, the degree of these goals in the body, how concentrated the medicine in healthy and sick tissues and the structural properties of the medicine binds.

These characteristics of AI-generated drugs could be tested in the so-called phase-0 studies, with ultra-lowering doses being used in patients with serious and lighter illness. This could help the researchers to identify optimal medication and at the same time reduce the costs of the current “test-and-lake” approach for clinical studies.

While AI alone may revolutionize the development of pharmaceuticals, it can help to fix the basic causes of why medication fails and optimize the lengthy process for admission.

This dux, Associate Dean for Research, Charles Walgreen Jr. Professor of Pharmaceutical and Pharmaceutical Sciences, University of Michigan and Christian Macedonia, extraordinary professor of pharmaceutical sciences, University of Michigan

This article will be released from the conversation under a Creative Commons license. Read the original article.

When China's Deepseek threatens to reduce Silicon Valley's AI monopoly, a European alliance has been created with an alternative to the global order of Tech.

You call your project Openeurollm. Like Deepseek, you want to develop open source language models of the next generation-but your agenda is very different. Your mission: European AI, which will promote digital managers and effective public services on the entire continent.

To support these goals, Openeeurollm builds a family of powerful, multilingual large language foundation models. The models will be available for commercial, industrial and public services.

Over 20 leading European research institutions, companies and high -performance computers (HPC) centers have entered the project. The leadership of her alliance is Jan Hajič, a renowned computer linguist at Charles University, the Czech Republic, and Peter Sarlin, co-founder of Silo Ai, the largest private Ai laboratory in Europe was acquired Last year by US chip maker AMD for $ 665 million.

Webinar: Promote the success of the SCALEUP success

Visit us on February 18 to promote a discussion about the important role of ecosystems in promoting startups and scale -ups and promoting a dynamic entrepreneurial landscape.

You are accompanied by a number of European tech lights. Among them are Alph AlphaThe leading light of the German AI sector, Finland CSC, which houses one of the world The most powerful supercomputers., and France's lights that recently became Europe's first listed genai companies.

Your alliance was supported by the European Commission. According to Sarlin, the initiative could be the Commission's largest AI project.

“What is unique about this initiative is that we bring many leading AI organizations in Europe together in a focused effort instead of having many small, fragmented projects,” he said TNW via e -mail.

“This concentrated approach must need Europe to build open European AI models that ultimately enable innovations on a scale.”

The project has a budget of € 52 million and a calculation of the engagement, which may have a greater monetary value, said Sarlin.

In addition to the financing of the Commission, Openeeurollm was supported by Step, an EU program to increase investments in strategic technologies.

The project also corresponds to the plans of the EU to strengthen the digital sovereignty of Europe, which is susceptible.

Europe's AI future

Since China and the USA are developing new AI skills at good speeds, Europe sees itself in the digital landscape of an uncertain future.

Openeurollm hopes to strengthen the position of the continent with a new digital infrastructure. The project has also undertaken to embed AI into European values ​​for democracy, transparency, openness and commitment of the community.

According to Openeeurollm, the models, software, data and evaluation will be completely open. You can also be able to make fine tuning and teaching votes for certain needs of industry and the public sector. In addition, alliance promises to preserve both linguistic and cultural diversity.

The plans come in test times for European technology. With the US and Chinese companies that provide new AI breakthroughs, fears grow that European companies, economies and even culture are threatened.

Sarlin wants Openeeurollm to bring the continent new hope.

“This is not about creating a chat bot for general purposes. It is about building the digital and AI infrastructure that enables European companies to be innovative with AI,” he said.

“Regardless of whether it is a healthcare company that develops specialized assistants for doctors or a bank who creates personalized financial services, you need AI models that are adapted to the context in which you are active and that you control and can have.

“This project is about giving European corporate instruments to create models and solutions in your languages ​​that you own and control.”

A German startup plans to start European EVs with a AI-driven brain.

The Sphere Energy has built the system to simulate the battery behavior. The company then predicts the lifespan of a power source in numerous scenarios, from driving styles to temperatures on the street.

According to Sphere, the findings reduce the battery test cycle by at least one year. The development of a car could now be “at least” completed twice as quickly.

Sphere introduces endless advantages: manufacturers will save millions, car prices decrease and innovations will increase to exponential rates.

Webinar: Promote the success of the SCALEUP success

Visit us on February 18 to promote a discussion about the important role of ecosystems in promoting startups and scale -ups and promoting a dynamic entrepreneurial landscape.

The co -founder of the startup, Lukas Lutz, said the plans were unprecedented.

“Nobody at the moment – not even Tesla – can appreciate the lifespan of your battery,” Lutz told TNW. “That will be really groundbreaking.”

A lifeline for European EVS?

Sphere presented the project at the last month at the IBM research Laboratory in Switzerland.

In a futuristic facility with a view of Lake Zurich, the startup introduced a KI brain called Batty.

BATY was initially trained over the years of the test data of over 1,000 batteries. Car manufacturers also mix their own information. The system then simulates the lifespan of a certain battery under different conditions.

Customers can test the effects of acceleration of the highways and crawling in mountains, apply fast and slow chargers, drive in the scenery and freeze winter. Each aspect affects the reduction of the battery.

The power of the system comes from the transformer architecture – the founding stone of today's large voice models (LLMS). But Spheres approach is not just based on text. The startup extends the area of ​​the model by integrating time cereal data. As a result, the system can simulate the behavior of a battery for years.

The approach gives the LLM paradigm a new turn. While a chatbot predicts the next best word, Batty will predict the next best data point.

Car companies were impressed by the results. According to Sphere, the majority of European manufacturers have already used the technology.

BATY could give EV manufacturers of the continent an important thrust that quickly lose their Chinese rivals the market share.

“The development of the battery is a great pain for them – and it shouldn't be,” said Lutz. “We really want to take the load away.”

But batteries are just the beginning of the ambitions of Sphere. The company provides itself with simulating endless energy applications, from electric boats to mains stores.

In addition to IBM, the startup also examines new levels of simulation of batteries.

“With these foundation -KI models, we intrinsically understand the atomic level,” said Lutz. “But we want to be subatomic-with quantum.”

The rapid spread of artificial intelligence asked people: Who is most likely in their daily life? Many assume that it is the technically experienced-The ones who understand how AI works-who are the most on taking over it.

Surprisingly, our new research (published in the Journal of Marketing) finds the opposite. People with fewer knowledge of AI are actually more open to the use of technology. We call this difference in the tendency to adoption the connection “lower literacy high susceptibility”.

This link is displayed in different groups, settings and even countries. For example, our analysis of data from the market research company Ipsos, which includes 27 countries, that people in nations with lower average AI alphabetization are more susceptible to the introduction of AI than people in countries with higher literacy.

Similarly, our survey among US students finds that those with less understanding of AI are more likely to indicate that they are used for tasks such as academic tasks.

The reason for this link is how AI is now doing tasks that we once thought that only people could do it. When AI creates a work of art, writes a warm answer or plays a musical instrument, it can feel almost magically – as if it were going into the human area.

Of course, AI has no human properties. A chat bot could create a sensitive reaction, but it doesn't feel sensitive. People with more technical knowledge of AI understand this.

You know how algorithms (sentences of mathematical rules used by computers to execute certain tasks), training data (used to improve the functionality of a AI system) and arithmetic models. This makes the technology less mysterious.

On the other hand, those with less understanding can see AI than magical and impressive. We suggest that this feeling of magic makes you more open to use AI tools.

Our studies show that this recovery connection, which is prospective with lower literacy, is strongest for the use of AI tools in areas that associated people with human characteristics, e.g. B. emotional support or advice. When it comes to tasks that do not emerge the same feeling of human characteristics as the analysis of test results, the pattern flips. People with higher AI alphabetization are more susceptible to these uses because they focus more on the efficiency of AI than on “magical” properties.

It's not about skills, fear or ethics

Interestingly, this connection remains between lower literacy and greater sensitivity, although people with lower AI alphabetization look at the probability of less capable, less ethical and even a bit scary. Her openness to AI seems to be due to her feeling of amazement at what she can do, despite this perceived disadvantages.

This knowledge provides new insights into the question of why people react so differently to the technologies. Some studies suggest that consumers prefer new technology to a phenomenon called “Algorithm value estimation”, while others have skepticism or “algorithm difference”. Our research indicates the perception of AIS “magic” as a key factor that shapes these reactions.

These findings are a challenge for political decision -makers and educators. The efforts to increase the AI ​​alphabetization could unintentionally dampen people's enthusiasm for the use of AI by making less magical. This creates a difficult balance between the support of people to understand AI and to keep them openly to their adoption.

In order to optimize AI's potential, companies, educators and political decision -makers have to increase this balance. If we understand how the perception of “magic” influences the openness of people for the AI, we can help to develop and use new products and services on AI-based AI that take into account the way people take into account , and help them to understand the advantages and risks of AI.

And ideally, this will happen without causing a loss of awe, which inspires many people to use this new technology.

Chiara Longoni, Associate Professor, Marketing and Social Science, Bocconi University; Gil Appel, Assistance Professor of Marketing, School of Business, George Washington University, and Stephanie Tully, Associate Professor of Marketing, USC Marshall School of Business, University of Southern California

This article will be released from the conversation under a Creative Commons license. Read the original article.