Jump directly to content

Ariel Flint Ashery: More is different – emergent social conventions and tipping points in AI populations

At RISE Learning Machines Seminar on August 27th, 2026, we have the pleasure to listen to Ariel Flint Ashery, City St George’s, University of London, give his talk: More is different – emergent social conventions and tipping points in AI populations.

This seminar is a collaboration between RISE and Climate AI Nordics – climateainordics.com.

Seminar Details

When: August 27th 2026, 15:00 CET
Where: Online via Zoom

Register here

Abstract

+

About the speaker

+

Olof Mogren

Principal Researcher
+46 73 023 56 09 Read more about Olof
Profile image

Contact Olof

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.
learning machines seminars
bg_1
Off
Artificial intelligence

How Nexans began turning production data into valuable insights

How Nexans began turning production data into valuable insights

How can a manufacturing company use its production data to reduce disruptions, improve quality and make more sustainable decisions?
At its Grimsås site, Nexans has transformed fragmented data into valuable insights.

"Our goal has not only been to focus on deviations, but also to receive feedback from digital solutions when we're doing things right," says Perparim Sylejmani at Nexans.

I Grimsås tillverkas allt från kraftkablar för elnät och energiförsörjning till kablage åt fordonsindustrin. Nexans Sweden har sett en kraftig efterfrågeökning mot bakgrund av energiomställningen, och för att möta den har företaget bland annat investerat i nya, moderna produktionslinjer. 

Full spårbarhet i produktionsflödet 

Med högtryck i produktionen ökar också kraven på stabil drift, kvalitet och spårbarhet. Här har digitalisering varit en nyckel. 

– Vi i Grimsås blev utvalda att vara pilot i koncernens digitala transformation. Där ingick att rulla ut ett nytt grundsystem för tillverkning, Manufacturing Execution System, som ger oss möjligheter att ha kontroll på tillverkningen, maskinerna och produktparametrar. Systemet ger full spårbarhet i flödet och standardiserar vårt sätt att producera. Vi har också installerat cirka 200 energimätare för att kunna länka förbrukning till specifika produkter, berättar Perparim Sylejmani, Digital Factory Leader på Nexans Sweden. 

Med en bra digital grund kunde nästa steg på resan tas. Nexans ville börja jobba med förebyggande och konditionsbaserat underhåll, och även bli bättre på att hantera och visualisera datamängderna som skapas i produktionen. 

– Mycket har handlat om att skapa en datakultur genom kompetensutveckling och förändringsledning. Vi behöver förstå vilken data som har betydelse för oss. Dels för att kunna fatta kortsiktiga beslut i produktionen, dels för att driva ett ständigt förbättringsarbete. Målet har varit att inte bara fokusera på avvikelser, utan också få feedback från digitala lösningar när vi gör rätt, säger Perparim Sylejmani.

För att fånga helheten kan vi inte bara ta in AI-experter, utan vi måste också involvera personer som forskar kring företagsutveckling, arbetssätt, hållbarhet och människans beteende.

Perparim Sylejmani, Digital Factory Leader, Nexans Sweden

Behov av AI-kompetens ledde till forskningssamarbete 

Nexans representanter förstod att svaren på färre driftstopp, jämnare kvalitet och mer effektiv produktion fanns i produktionsdatan. Redan under förstudierna av den digitala transformationen insåg Perparim Sylejmani att Nexans behövde kroka arm med forskare och oberoende experter för att kunna nyttja produktionsdata och använda AI på smartast möjliga sätt för att gräva fram svaren. 

– Det finns fantastiska förutsättningar för svensk industri att dra nytta av resurser inom RISE, universitetsvärlden och Vinnova. Att aktivt delta i forskningsprojekt borde alla företag göra. Vi är snart inne på vårt tredje projekt tillsammans med RISE och Chalmers, berättar Perparim Sylejmani och fortsätter: 

– Vi bestämde oss tidigt för att ha en bredd i projekten. För att fånga helheten kan vi inte bara ta in AI-experter, utan vi måste också involvera personer som forskar kring företagsutveckling, arbetssätt, hållbarhet och människans beteende. Vi lär oss oerhört mycket genom de här tvärfunktionella mötena. För oss är det en helt fantastisk chans att på veckobasis sitta ner tillsammans med forskare från RISE, och att forskarna har kommit till oss i Grimsås för att förstå vår verklighet. 

RISE har bidragit med kompetens inom AI, dataanalys och datadrivna arbetssätt, och utvecklat metodik för att analysera och söka i produktionsdata. 

Expertstöd längs hela förändringsresan 

Sedan starten av samarbetet 2021, har forskare och experter från RISE och Chalmers arbetat i tätt samarbete med nyckelpersoner inom digitalisering, produktion, underhåll och kvalitet på Nexans. RISE:s expertis har funnits med hela vägen – från utveckling av analysverktyg, framtagande av analysmetodik och inte minst förankring i organisationen. 

– I början låg fokus på att förstå vilka tillgängliga data som finns och integrera dem för att förstå dess sammanhang, säger Camilla Munther. 

Därefter följde utvecklingen av en första version av verktyget. Denna testades och har vidareutvecklats tillsammans med användare i fabriken. 

– Det handlar om hur data kan struktureras och visualiseras, och användas för att identifiera variationer och avvikelser. 

Därtill gav RISE stöd i själva förändringsarbetet på Nexans, genom att förankra det datadrivna arbetssättet hos fabriksledningen och koppla analysen till verksamhetens mål. Sammantaget gav det Nexans ett smartare sätt att använda den data som ändå genereras. 

– Vi bidrog också med kompetensutveckling, bland annat genom dialog med operatörer och produktionsteam för att möjliggöra användning av verktyget i det dagliga arbetet. Vi har arbetat med en produktionslinje - nu planeras utbildning för att skala upp det till andra delar av fabriken. 

– Förutom tillgång till oberoende experter får företagen även chans att utbyta erfarenheter med varandra. Det som är vedertaget i en bransch kanske är helt nytt för en annan, så här finns det också värden att hämta, säger Camilla Munther. 

Perparim, vad har kommit på plats hos Nexans efter projekten? 

– Tillsammans har vi utvecklat ett system för likhetssökningar, SIFT. Nu kan vi koppla ihop tidseriedata (parametrar som temperatur, energiförbrukning och vibrationer i varje given sekund) med metadata från vårt Manufacturing Execution System. Metadatan utgörs av information om maskininställningar, tidigare kvalitetsmätningar och så vidare, förklarar Perparim Sylejmani.

Bilden visar likhetssökningsvyn i SIFT-verktyget utformat för visualisering, jämförelse och bedömning av historiska tillverkningsordrar. I vyn visas tidsseriedata för den valda tillverkningsordern, tillhörande metadata samt en lista av de mest liknande tillverkningsordrarna.

Kombinationen av data synliggör vilka förutsättningar som krävs för optimal tillverkning av specifika produkter. Den kan också fungera som ett facit vid problem och användas för att utbilda nya operatörer. 

– Med det här verktyget kan vi nå ett annat djup i rotorsaksanalysen. Vi kan också följa råvaruförbrukning i realtid, i gram per meter och kilometer. Det här är lika mycket en hållbarhetssatsning som en digitaliseringsresa, säger Perparim Sylejmani. 

Förhoppningen hos projektdeltagarna är att även andra företag ska implementera verktygen och arbetssätten, som ett led i digitaliseringen av svensk industri. Produktionsdata finns det gott om, men utnyttjandegraden varierar. 

Från datasilos och strutfotboll – till överblick och lagspel 

Enligt Camilla Munther är det vanligt förekommande att tillverkningsindustrins data ligger i silos. Inte bara rent tekniskt, i olika informationssystem, utan också sett till människorna i organisationen. 

– Vi använder ibland strutfotboll som en liknelse. Fabrikens olika funktioner springer på sina bollar, och försöker få ordning på produktionsstörningar med de egna projekten i fokus. Ingen har riktigt koll på vem man har bredvid sig, eller ser omfattningen av störningen. Att göra en liknande förflyttning som Nexans handlar mycket om att sluta spela strutfotboll och börja spela som ett sammansvetsat lag, säger Camilla Munther. 

– Det här är särskilt kopplat till att gå från data i silos till gemensamma, verksamhetsnära insikter hela vägen - från data till insikt och från insikt till åtgärd, säger Camilla Munther. 

Vad tänker du att andra kan lära av Nexans resa? 

– Att våga testa! Jag upplever att Nexans är väldigt villiga att prova nya saker. De är nyfikna och på hugget. Det har resulterat i digitala verktyg som formar Nexans arbetssätt, så att dessa blir hållbara över tid. Att få återkommande guidning och tillgång till forskning genom oss innebär att företaget hela tiden kan utveckla den delen. 

Det sammansvetsade lagspelet kommer att utforskas vidare i ett nytt projekt. Perparim Sylejmani vill skapa förutsättningar för en integration mellan SIFT och operatören på produktionsgolvet. 

– Operatören kan säga att "jag vet vad som hände vid det tillfället då vi fick en avvikelse. Jag doserade material, det var därför det blev ett kort avbrott". Vi kan se trender genom likhetssökningen, men det är operatörerna som är experter på produktionsprocessen och "sina" maskiner. Min förhoppning är att vi ska kunna nyttja operatörens sinnen i kombination med verktyget. Det tror jag är framtiden, säger han.

About Nexans

Nexans Sweden is headquartered in Grimsås, in the municipality of Tranemo, and is part of the global Nexans Group, headquartered in Paris, France. The Grimsås site manufactures power and installation cables, alarm and security cables, automotive cables, as well as fibre optic and optical cables. It also has its own research and development department and an in-house laboratory.

About the partnership

Nexans Sweden has been collaborating with RISE and Chalmers since 2021 through a series of sub-projects funded by Vinnova. The aim is to develop tools that can benefit other manufacturing companies and help accelerate the digitalisation of Swedish industry.

SIFT – Similarity Search for Time Series Data is supported by Vinnova through the Advanced Digitalisation programme.

Project partners: RISE, Chalmers, Nexans, Nord-Lock and IFM Electronic.

Production and manufacturing Sekundär områdes navigation:
Artificial intelligence
Digitalisation

How to move from an AI pilot project to genuine value creation

AI

Most organisations have started using AI. The challenge is no longer to start using the technology, but to ensure it creates real value. The problem is that AI often remains confined to pilot projects and technology initiatives, rather than transforming working methods and processes, which is where costs, revenues and the business itself actually lie.

AI is currently used for a variety of purposes, including customer service, analytics, product development and automation. However, international surveys show that over half of companies struggle to identify a clear financial return on their AI investments.

We need to progress beyond the initial phase, which involves smaller innovation groups.

This is not particularly surprising, according to Martin Körling, head of the data analysis unit at RISE.

‘Setting up a pilot project alongside the main business isn’t difficult. However, it won't generate any new revenue or result in any change in costs for the main business,” he says.

Many organisations begin by establishing small AI teams or innovation groups to explore the technology. This is a natural first step.

‘But you have to get past that.' In order to move forward, you have to integrate this expertise into your operations,' says Martin Körling.

He also believes that this is where many AI initiatives get stuck. Value is only created when AI influences an organisation's working methods and core processes.

‘It is the managers in the core business who need to start thinking about how AI can be used in their processes,’ says Martin Körling. That’s when things will start to happen,” says Martin Körling.

Integrating AI into core business operations requires more than just new tools.

However, transitioning AI from pilot projects to core operations is a complex process that requires more than just purchasing licences for new tools.

According to Martin Körling, organisations need to understand how developments in AI are changing their operations. They must also have the courage to challenge existing processes, identify which parts can be automated and grasp how their products and services create added value.

At the same time, the standard of data, security and infrastructure must be high enough. To scale up the use of AI, organisations also need to understand the risks and regulatory framework, and learn how to use the technology responsibly.

The ease with which AI can be implemented does not necessarily depend on the size of a company. The difference is often greater between organisations where data is already available and those that first need to digitise their processes.

‘If you run a sawmill and haven’t started collecting data, you might need to begin by taking photographs of your logs. Based on those pictures, you can then do lots of interesting things,” says Martin Körling.

In order to move forward, you have to integrate this expertise into your business.

Value and return are not the same thing.

The discussion about AI often centres on return on investment (ROI). However, value and return do not always align in the short term.

‘You can improve the product by doing things in a new way, but you still can’t charge more for it. Even if it isn’t possible to quantify it initially, it can still be incredibly important,' says Martin Körling.

Nowadays, AI is often considered separately – as a project, a pilot scheme or a specific function.

Martin Körling believes that, in the long run, this division will disappear.

He compares this development to how the internet has transformed businesses and organisations.

‘When the internet first came along, it was seen as something separate. Now, however, it’s built into everything, he says.

Spotify is a prime example of what the future might look like. Although recommendations, playlists and other features rely heavily on AI, most users don’t give the technology behind them a second thought.

‘There is no “AI” button in Spotify. It’s deeply integrated into the product, which you simply experience as being better: more precise and in line with your tastes and expectations,” says Martin Körling.

Assessing the maturity of AI can help to identify the next steps.

For many organisations, carrying out an AI maturity assessment can be the first step towards understanding their current situation and identifying their next steps.

According to Martin Körling, understanding the business is as important as understanding the technology.

Among other things, RISE helps organisations understand the impact of AI on their operations, identify new working methods and develop prototypes to facilitate the testing of ideas in practice.

At the same time, RISE's expertise covers everything from industry and energy to forestry. This enables AI issues to be linked to the specific operations in which the technology is to be used.

“We can talk about AI technology itself. But we can also discuss how it affects businesses,” says Martin Körling.

Artificial intelligence Sekundär områdes navigation: Digitalisation

AI support can make public information more accessible to a wider audience

A person wearing a brown jacket holds up a smartphone with the camera side facing the viewer, photographed outdoors.

Convoluted official documents and local authority websites can make it difficult for people with intellectual disabilities to understand their rights and take control of their own lives. Could new digital tools and user-driven service design be part of the solution? This is what Jönköping Municipality set out to discover.

In many ways, digitalisation has made it easier for information to flow between local authorities and residents. These days, we no longer need to keep track of paper forms for applying for a nursery place or leaflets stating when rubbish is collected – all that information is available on the local authority’s website.

However, accessing information can be challenging for people with intellectual disabilities. A report by the Swedish Post and Telecom Authority titled "Swedes with Disabilities and the Internet 2025" shows that people with disabilities find it significantly more difficult to use the internet than others. Many experience recurring problems when searching for information, navigating websites, and understanding content. The report concludes that digital services are often poorly adapted to users’ needs, which leads to lower uptake of essential public services and reduced participation in society.

Jönköping Municipality has been addressing this issue actively for several years. In 2025, the Social Services Department appointed Fredrik Härnström as a support teacher specialising in digital development and inclusion within the disability support division. He is responsible for training and assisting both residents and staff in using digital tools to make everyday life easier for people with intellectual disabilities.

– Under the Social Services Act, we must ensure that everyone who needs support receives it so they can live like other residents of the municipality. Our work with RISE has been about taking steps in that direction, says Fredrik Härnström.

Under the Social Services Act, we must ensure that everyone who needs support receives it so they can live like other residents of the municipality. Our work with RISE has been about taking steps in that direction.

Fredrik Härnström

User-driven service design: a winning formula.

Jönköping Municipality has collaborated with Skellefteå and Gothenburg municipalities on a Vinnova-funded project with RISE. The aim of the project was to develop a prototype AI support tool for people with intellectual disabilities. This involves a chatbot designed to act as an accessibility filter for public information. Using generative AI, users will be able to have text read aloud to them, or have it simplified or explained using visual aids.

The target group was involved from an early stage. Fredrik Härnström, a former teacher at a special needs upper secondary school who is now in regular contact with Jönköping’s day centres, had no difficulty finding a dozen people willing to take part in the project.

– During the first round of interviews, we asked the participants how they currently use the internet. We asked them how they search for information and what they find challenging. One question was about how often they visit the local council’s website. I suppose only a handful of them had actually visited it; it wasn’t exactly the most popular website, says Fredrik Härnström, chuckling.

The secure RISE platform provided a solid foundation for the prototype

Jonas Ek, project manager at RISE, provides further insight into the technology behind the prototype.

– We have developed a core platform called RISE GPT. It is a language model platform with essentially the same functions as ChatGPT, except that no information is stored outside Sweden or Europe. RISE GPT serves as a learning environment for local authorities, regions, and government agencies that are part of our AI partnership, explains Jonas Ek.

– We also used this platform as the basis for the project in collaboration with Jönköping, Skellefteå and Gothenburg. We also opted to work with RAG technology. This involves gathering information — in this case, data on day-to-day activities in the municipality of Skellefteå — and storing it on a server that the AI tool can search. While this technology reduces the risk of "hallucinations", the responses can sometimes be too narrow or dry. It was a challenge we had to work hard to overcome.

Another challenge was that the target group was as diverse as any other. Some were highly experienced app users who wanted the tool to offer a full range of options. Others just wanted a single button. The local authorities developed personas based on these different needs. (A persona is a fictional yet realistic profile of an individual who represents a typical target group or target group segment.) With these in mind, the developers created a 'prompt-Lego' system comprising various Lego bricks (instructions for the AI) that could be combined to customise the presentation of information to different users.

– Once we had a prototype to test, we visited the participants in our day-to-day activities again. That’s when a number of bugs and problems came to light, which proved useful for the next stage, says Fredrik Härnström.

Relevant experience in the public sector as a whole

The project's primary aim was to learn from working inclusively with AI tools for people with disabilities, rather than to produce a finished application. Jönköping, Skellefteå and Gothenburg have gained a deeper understanding of generative AI and how AI tools can be developed iteratively. They have also learned what to consider when drawing up specifications for the procurement of AI solutions. These experiences could also benefit other local authorities.

The prototype that has been developed can now be refined further by each local authority to create a fully-fledged support tool. Jonas Ek explains that the project and prototype presentations given at various trade fairs have generated significant interest. This bodes well for the project’s future.

– We can see that there are good opportunities to develop this work further. I believe that, as AI becomes an integral part of standard business systems, we will see a "ketchup effect" across Sweden’s local authorities in the future, with more local authorities taking part in this kind of collaboration, says Jonas Ek.

I believe that, as AI becomes an integral part of standard business systems, we will see a 'ketchup effect' across Sweden’s local authorities in the future, with more local authorities taking part in this kind of collaboration.

Jonas Ek

Jonas Ek

Projektledare
+46 73 071 52 27 Read more about Jonas
Profile image

Contact Jonas

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.
Artificial intelligence Sekundär områdes navigation:
Digitalisation
Digital infrastructure

Advanced Material Characterisation Seminar Series

Industries face increasing pressure to develop materials that are lighter, stronger and more sustainable – without compromising durability, functionality or safety.
This seminar series explores how advanced characterisation techniques, such as X-ray and neutron methods, reveal what happens inside materials in real time and under realistic conditions.
From recycled materials and green chemistry to catalysts and pharmaceuticals, the series shows how deeper insight supports better decisions in R&D and innovation.

Harnessing neutron scattering techniques for innovative product and material development: User case studies and applications

Randi Nordström, RISE

Small-angle x-ray scattering for pharmaceutical industry challenges

Christopher Söderberg, RISE

In-situ analysis of cast metals using mechanical behaviours x-ray tomography 3DXRD

Torsten Sjögren, RISE

Transmission electron microscopy and sample preparation

Annika Altskär, RISE

Torben Nilsson Pingel, RISE

X-ray nano diffraction imaging for soft and hard materials

Shun Yu, RISE

3D CT images from 2D x ray scanning with Physics AI

Benedikt Neyses, RISE

CT at RISE

Malte Wallmeier, RISE

AI for segmentation of x ray tomography images

Sandra Baman, RISE

Residual stress analysis of materials at RISE

Jonas Holmberg, RISE

Structural location, in-situ conductivity and hydration studies of proton conductors using neutron powder diffraction

Istaq Ahmed, SEEL

Using neutrons in biologics

Sara Malekkhaiat Häffner, RISE

Modification of a microwave oven for time-resolved synchrotron X-ray microtomography of bread baking

Sven Isaksson, RISE

Micro-x-ray fluorescence microscopy and application to concrete durability

Gilles Plusquellec, RISE

3D and 4D imaging of material microstructure -focus on battery mechanics

Professor Stephen Hall, Division of Solid Mechanics, Lund University, Sweden

Niklas Lorén

Projektledare
+46 10 516 66 14 Read more about Niklas
Profile image

Contact Niklas

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.

Fredrik Holmgren Holm

Innovations- och processledare
+46 10 516 69 14 Read more about Fredrik
Profile image

Contact Fredrik

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.
Simple X webinar
bg_1
Off
Artificial intelligence

Markus Reichstein: Machine Learning for Earth System Science

At RISE Learning Machines Seminar on June 25th 2026 we have the pleasure to listen to Markus Reichstein, Max-Planck-Institute for Biogeochemistry and ELLIS Unit Jena, give his talk: Understanding the complex Earth system with Machine Learning and Hybrid Modelling.

This seminar is a collaboration between RISE and Climate AI Nordics – climateainordics.com.

Seminar Details

When: June 25th 2026, 15:00 CET
Where: Online via Zoom

Register here

Abstract

+

About the speaker

+

Olof Mogren

Principal Researcher
+46 73 023 56 09 Read more about Olof
Profile image

Contact Olof

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.
learning machines seminars
bg_1
Off
Artificial intelligence

Gustau Camps-Valls: A Critical Look at Explainable AI

At RISE Learning Machines Seminar on June 11th 2026 we have the pleasure to listen to Gustau Camps-Valls, University of Valencia, give his talk: A Critical Look at Explainable AI.

This seminar is a collaboration between RISE and Climate AI Nordics – climateainordics.com.

Seminar Details

When: June 11th 2026, 15:00 CET
Where: Online via Zoom

Register here

Abstract

+

About the speaker

+

Olof Mogren

Principal Researcher
+46 73 023 56 09 Read more about Olof
Profile image

Contact Olof

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.
Gustau Camps-Valls, University of Valencia
bg_1
Off
Artificial intelligence

Ariel Flint Ashery: More is different – emergent social conventions and tipping points in AI populations

At RISE Learning Machines Seminar on June 4th, 2026, we have the pleasure to listen to Ariel Flint Ashery, City St George’s, University of London, give his talk: More is different – emergent social conventions and tipping points in AI populations

This seminar is a collaboration between RISE and Climate AI Nordics – climateainordics.com.

Seminar Details:

When: June 4th, 2026, 15:00 CET
Where: Online via Zoom

Register here

Abstract

+

About the speaker

+

Olof Mogren

Principal Researcher
+46 73 023 56 09 Read more about Olof
Profile image

Contact Olof

CAPTCHA

* Mandatory 

By submitting the form, RISE will process your personal data.
Ariel Flint Ashery, City St George’s, University of London
bg_1
Off
Artificial intelligence