AI-greetings from Tallinn – Insights from the FAIR's Project Manager at New Nordics AI Week

The AI/ML Forum in Tallinn in September launched the New Nordics Week 2024, featuring a series of AI-focused events, including a high-performance computing session organized by FAIR and Finnish EDIHs. Read on about FAIR’s Manager Solja Sulkunen’s Global AI Insights to deliver insights and foresights from the field of AI, XR and HPC around the world on a very concrete level; what is being discussed right now, where is AI heading? Sulkunen shares her key learnings and thoughts provoked by the New Nordic AI Week

 

FAIR's project Manager Solja Sulkunen and EU flag on the background

27.9.2024

Nauha, jossa FAIR:in yhteistyökumppanien logot.

3 examples of Real life use cases of AI in enterprises

Whether we talk about drug design and development or sales automation platform, they both respond to a very specific task to be improved. This is where AI can be used in product development. For example, in drug development, the need is to cut down the time used for designing a new drug. The benefit is that AI powered design allows us to generate and test different variations of molechyle combinations in significantly shorter time than a human ever could, even up to 50% ­- meaning tens of months in time. The next interesting thing to research is, whether these AI designed drugs will actually also work better than human designed drugs. By shortening this time, the time to respond to medical conditions and enter the markets will also be shortened. In sales automatization the process to look into is the time closing sales, by handling multiple cases simultaneously and receiving AI-based suggestions when and how to connect a client to ensure closing the deal.

AI can be also utilized to support a company in executing its core business, by using process optimization. Let’s take as an example a company selling and setting up and solar panels, who wants to boost up their sales processes through faster handling with equal quality. The company could be a customer of the AI powered sales platform company, but also large language models could be the answer to many different internal and external company needs. Internal AI applications include automated memo and case documentation, summarizing reports, extracting information from large datasets, generative AI for image/video editing, and aiding software/resource management. IBM’s representative estimates time savings of 20-50% based on task complexity.

What seems to be expected to happen in somewhat near future, is that the use of AI as agents in everyday work will increase. Professionals will transfer their routine tasks to AI, having them respond to emails, scheduling meetings, making travel arrangement, preparing meeting agendas, memos and action points, and writing reports, as an example, independently. However, a key question here is, how will we utilize the capacity and resources saved in an intelligent and meaningful way? 

Change Management

In New Nordics AI Week there was a quite big consensus that the technology is good enough to be utilized as we speak, even if rapid development is expected. It was highlighted that the honeymoon phase of piloting with AI is now over, and that now we need to figure out how to tackle the elements hindering the deployment and scaling of AI, such as complience, regulation and by-in of the actual users.

Once the groundwork is set, managers should present a clear, actionable plan.

 Change management takes time, and to foster the adoption of new technologies and behaviors, managers must begin the process well before introducing new solutions. Understanding the criticality of the technologies, development and change from one’s own perspective is crucial. Once the groundwork is set, managers should present a clear, actionable plan and offer opportunities for feedback. The plan can then be adjusted based on insights from those affected by the change. Regular meetings with employees before, during, and after the process are essential for gathering feedback and managing emotions.

As a rule of thumb, every little change in how we do things require individual learning of new habits, and changing even one habit is hard and timely. 

Open source & Decentralized AI 

Open source and decentralized AI are two pivotal concepts shaping the future of technology. Open source refers to software whose source code is freely available to the public, allowing anyone to inspect, modify, and enhance it. Today, this often means hosting code on platforms like GitHub, though some argue this centralization isn’t ideal.

Decentralized AI, on the other hand, envisions a model where AI development is distributed across multiple entities, with collaborative decision-making. Critics of open source in defense technology worry that openly available code could fall into the wrong hands, undermining democratic values. However, proponents argue that innovation thrives on transparency, as understanding past developments is crucial for future advancements. They believe that bad actors will find ways to access solutions regardless of whether the code is open or not. 

Security, privacy and regulation 

Let’s imagine we have found a solution that responds to our business driven need, we have the data to feed the solution with and we also have the whole company onboard in the change. What was discussed on top of all this was the security and regulation related aspects of deploying AI solutions. Quite quickly cyber security discussion seems to turn into a discussion about defense – what would this mean from a small and medium sized company’s perspective?

To strike the right balance, continuous dialogue is essential between those responsible for defining regulations and the stakeholders actively engaged in technological development.

When addressing security, two key aspects must be introduced immediately: business-critical information and data provider security. SMEs experimenting with AI solutions need to ensure that all employees understand what constitutes critical information and how open solutions, like large language models, function. Key questions include: What information can be shared and where? Who has access to the database? What are the motivations of those with access, and how will they use the data? Following GDPR guidelines is crucial for privacy protection, but how can you prove your solution secures personal data and that your team knows how to handle it responsibly?

The overall purpose of regulation is generally acknowledged and accepted, but the extent of its application remains a critical question. Excessive regulation can unintentionally undermine the EU’s goals for technological sovereignty and stifle entrepreneurial innovation. To strike the right balance, continuous dialogue is essential between those responsible for defining, designing, and enforcing regulations and the stakeholders actively engaged in technological development.

Advice from FAIR’s project manager

When ever condidering deploying AI for business benefits, it is advicable to have the following key documentation ready; 

  • Data management and governance plan
  • AI strategy 
  • Continuously updated documentation on the security level and according to relevant regulation of your solution 

 

If you want to discuss more about the opportunitites of AI and take action in it’s deployment, do not hesitate to contact FAIR through registering your company to our ecosystem, and we will be in touch with you. 

That’s a wrap from Norway this time – stay tuned for the next tech-news splash across Europe!
Solja Sulkunen, Project Manager FAIR

White logo of Finnish AI Region (FAIR EDIH). In is written FAIR - FINNISH AI REGION, EDIH
Euroopan unionin osarahoittama logo

Finnish AI Region
2022-2025.
Media contacts