NeuralTrust at The Barcelona Deep Tech Summit

The 5th of November, we showcased our software solution for companies looking to implement and scale Generative AI safely at The Barcelona Deep Tech Summit. At both our stand and on stage, we had the opportunity to share insights on managing security risks in the rapidly evolving landscape of generative AI.

What is The Barcelona Deep Tech Summit?

The Barcelona Deep Tech Summit has emerged as a central hub for the startup ecosystem, particularly in the fields of scientific innovation and technology transfer. This event is designed as a dynamic platform where members of the Deep Tech community—start-ups, researchers, investors, and industry leaders—can connect, share insights, and explore new opportunities for collaboration. By fostering these connections, the summit aims to build a critical mass that will propel the Deep Tech industry forward, making it a major driver of economic and technological progress.

Deep Tech represents a sector focused on creating breakthrough technologies with the potential to reshape the future. This industry addresses complex global challenges through advancements in areas like artificial intelligence, biotechnology, quantum computing, and sustainable energy solutions. The technologies developed within Deep Tech are poised to redefine how we live, work, and interact, positioning the sector as a key force in shaping tomorrow’s world.

The summit offered an excellent platform to engage with forward-thinking companies and professionals dedicated to advancing deep tech. It was an inspiring experience to be part of meaningful discussions on AI, security, and innovation, and to connect with other remarkable companies committed to shaping the future of technology:

Highlights of The Barcelona Deep Tech Summit 2024

The third edition of the Barcelona Deep Tech Summit saw record participation, with over 2,500 attendees—nearly double the previous year. This year’s event featured more than 70 experts and showcased 47 start-ups tackling critical societal issues across fields like biomedicine, quantum technology, energy transition, sustainable construction, and robotics.

Jordi Valls, Barcelona’s Deputy Mayor for the Economy, highlighted the summit’s role in fostering a high-value economy that directly impacts public well-being, emphasizing the need for strong public-private collaboration and networks to accelerate the deployment of Deep Tech solutions in society.

Organized by Barcelona Activa in partnership with five Catalan universities and supported by the Government of Catalonia and the Ministry of Industry and Tourism, the summit welcomed new strategic partners this year, including Aldea Ventures, IESE Business School, Barcelona Supercomputing Center, Mobile World Capital Barcelona, and Tech Barcelona. For the first time, the event was held within the Smart City Expo World Congress.

On the first day of the Barcelona Deep Tech Summit, Cristian Canton, Meta’s Head of Responsible AI, highlighted the enduring importance of responsible AI. He noted that responsible AI has evolved beyond academic research, requiring a collective effort from all sectors of society—from industry and governments establishing clear policies to users and communities. Canton emphasized the need for dedicated teams to proactively address potential negative impacts of AI before technologies are released to the market.

Rapid Growth of the Deep Tech Sector in Catalonia and Spain

The Barcelona Deep Tech Summit highlighted new data on the expansion of the Deep Tech sector. The latest Deep Tech Start-up Ecosystem Analysis by Acció shows that Catalonia now hosts 340 Deep Tech start-ups—a 6% increase from last year—employing 2,735 people and generating €166 million in revenue. Barcelona leads Europe in Deep Tech funding, securing €544 million from 2019 to 2024, surpassing cities like Copenhagen, Madrid, and Amsterdam.

According to Mobile World Capital Barcelona, Spain has 1,210 Deep Tech companies, with Catalonia and Madrid hosting 55% of them. These companies generate €2 billion in revenue and create 12,200 jobs. They focus primarily on sectors such as health (35.6%), ICT (17%), industry and manufacturing (12.4%), and energy and sustainability (12.4%).

NeuralTrust: Security in the age of Generative AI

At NeuralTrust, we were among the select AI-specialized companies invited to participate in the Barcelona Deep Tech Summit. On the floor, we shared insights on cybersecurity challenges in the age of Generative AI. Here are the main points from our presentation, titled Security in the Age of Generative AI.

Generative AI has marked the start of a new era, with rapid adoption and transformative potential across industries. However, along with these advancements come critical security risks. While companies worldwide are benefiting from GenAI by reducing costs and increasing revenue, many overlook the essential cybersecurity frameworks required to manage this powerful technology effectively. According to Boston Consulting Group, over 40% of companies using GenAI lack a defined security strategy, leaving them vulnerable to significant threats.

The conversational capabilities of AI assistants powered by LLMs are impressive, but they also bring unique risks, from data leaks to model manipulation. Misconfigured AI systems can expose sensitive information, harm brand reputation, and erode customer trust, as seen in recent cases with companies like Samsung and Air Canada.

The main security challenges with Generative AI stem from its highly flexible, "creative" nature. LLMs, trained on vast data, respond in a zero-shot manner, meaning they can answer questions without prior context. However, this adaptability also makes them susceptible to Prompt Hacks—techniques that manipulate responses to reveal confidential or inappropriate information. Such risks include unauthorized access to sensitive topics, competitor mentions, and data leaks. With little effort, malicious users can provoke concerning responses, including political content, inappropriate tones, and even free product offers that could lead to legal and reputational problems.

Here are some critical vulnerabilities companies must consider when deploying GenAI systems:

• Prompt Hacks: Malicious prompts can manipulate LLM responses and behavior.

• Functional Failures: Models can produce incorrect, incoherent, or off-topic responses.

• Privacy Failures: LLMs may inadvertently disclose internal company data or personal information.

• Impartiality Failures: Chatbots may produce biased or discriminatory responses.

• Availability Failures: Models can be overloaded, leading to service disruptions and resource waste.

At NeuralTrust, we tackle these vulnerabilities with a unified platform solution that includes:

• Red Teaming: Controlled attacks to test and strengthen model security.

• Firewall: A robust database to block critical attack routes.

• Monitoring: Real-time software to track chatbot performance and detect incidents promptly.

• Tracing: Ensures compliance by maintaining a log of chatbot interactions to meet legal standards.

• Analytics: In-depth statistics on user interactions, aiding model improvement and user behavior analysis. Every day, we partner with major companies interested in or actively deploying Generative AI. It's essential to integrate cybersecurity considerations from the outset to avoid potential pitfalls. Without this foresight, companies risk facing the very worst of these scenarios.

Whether you saw us at the Barcelona Deep Tech Summit or if you want to find out more about the events we assist, don´t miss all the updates on our social media channels and blog.