Artificial Intelligence and Cybersecurity. What new threats should we prepare for?
Discover how AI-driven attacks like deepfakes and GPT-3 impersonations are transforming cybersecurity risks. Learn strategies to protect your business from the next generation of cyber threats.

AI Cyberattacks: The Next Generation of Threats
By Josep Guitart Pardellans (Investment Analyst, Sabadell Venture Capital)
and Fernando Mateus (CEO, Kymatio)
Update – May 3, 2021
Recent headlines like “Several European politicians are deceived through a video call with a deepfake” and “Deepfakes are already wreaking havoc on European diplomacy” show that Artificial Intelligence is reshaping the cyber threat landscape.
Source: The Guardian
AI and Human-Machine Interaction
To succeed long-term, companies must embrace smart automation that integrates Artificial Intelligence. In cybersecurity, AI-driven attacks will change the rules for both criminals and victims.
Examples:
- Google Duplex – An AI system capable of making restaurant reservations by phone, demonstrating natural human-machine interaction.
- OpenAI GPT-3 – A language model with 175 billion parameters, capable of generating code, designs, political discussions, and even news articles indistinguishable from those written by professional journalists.
Despite its hype, GPT-3’s ability to mimic trusted communication creates serious risks in business environments.
Implications for Cybersecurity
- Sophisticated Impersonation – AI could impersonate executives, partners, or trusted entities to deliver phishing, vishing, or smishing attacks.
- Increased Internal Risks – Internal security incidents have grown 47% in the past year, with 85% linked to negligence or social engineering.
- High Financial Impact – Average costs exceed €7M per incident in large companies, and €40K in SMEs — where 60% close within 6 months of a major incident.
Attack technology is evolving faster than employee training, making scams increasingly harder to detect.
Preparing for the AI-Driven Threat Era
To counter these risks, companies must:
- Integrate the human factor into their cybersecurity strategy.
- Adopt human cyber risk management tools to proactively detect and respond to threats.
- Continuously train employees to recognize even the most sophisticated scams.
Kymatio acts as a cyber companion, supporting employees and organizations with personalized training and risk monitoring to strengthen cyber resilience.
📩 Contact Kymatio to prepare your team for the next generation of AI-driven cyberattacks.
Sources: OpenAI · Sam Altman on GPT-3 · Research Paper · Citibeats Article