Hacks of the Future and The Role of Humans and Hackers

The talk opens with ChatGPT's rapid uptake, viewed from a hacker's perspective. The speaker, who leads market-area security at Ericsson, runs a live demonstration: the chatbot is coaxed into writing a phishing email, file-stealing code, and Ducky Script that pulls a saved Wi-Fi password and emails it out. Plain-English prompts, refined over several tries, were enough to produce working tooling.

The middle section looks ahead to AI-driven reconnaissance and botnets, password cracking aided by adversarial networks, and deepfakes, including a voice-cloned executive that reportedly authorised a large bank payment. Shadow APIs, open source dependencies, browsers, cloud services and supply chains are presented as growing entry points. Citing several industry studies, the speaker argues that human error drives most breaches, and discusses decision fatigue, phishing susceptibility and attacker motives.

The closing points were to remember that AI-driven hacking has arrived, that APIs and open source libraries need proper inventories, and that people remain the weakest link. Practical measures included enforcing MFA and teaching staff security through visible role models. The Q&A covered executive accountability and whether AI can be limited, with the speaker doubting that regulation will arrive quickly.