Thousands of cybersecurity professionals gathered at San Francisco’s Moscone Center for RSAC 2025, where one topic dominated every conversation: artificial intelligence. The event buzzed with a mix of optimism for AI’s potential and serious questions about its real-world effectiveness. From packed keynotes to the expo floor, the industry grappled with how AI is reshaping the fight against cyber threats, leaving many to wonder what is truly useful and what is just hype.
The Great AI Debate: Separating Hype from Reality
For veterans like Derek Manky from Fortinet, the shift was undeniable. After 15 years of attending RSAC, he noted that AI was no longer just a buzzword on a slide but the undercurrent of nearly every discussion. This sentiment was shared across the conference, as the focus moved beyond initial excitement to practical application.
The conversation has matured. Tim Mackey of BlackDuck pointed out that the industry’s new challenge is to separate genuinely useful AI tools from the “smoke and mirrors.” This year, talk of more autonomous “agentic AI” became common. As AvePoint’s Tim Boettcher explained, “Agentic AI isn’t just ‘smarter’ AI — it does stuff on its own.” This capability presents both powerful defensive opportunities and entirely new attack surfaces for bad actors to exploit.
Evidence of this push was everywhere. Tech giant Cisco showcased new AI integrations in its security platforms, while Terra Security, a startup less than a year old, announced it had already raised $8 million to develop its own agentic AI system.
Are AI Defenses Keeping up with AI-Powered Attacks?
Amid the excitement, a strong sense of caution prevailed. Many experts questioned whether the industry’s defensive AI tools are truly ready for prime time. Tony Anscombe from ESET was blunt, stating that defenders are still playing catch-up. He voiced concerns that attackers are already using AI effectively to create sophisticated social engineering campaigns and smarter malware.
Anscombe argued that too many security vendors are marketing their “AI defenses” without providing clear evidence of their effectiveness. This leaves many security leaders asking the “million-dollar question,” as Michael DeBolt of Intel471 put it. DeBolt spent his time at RSAC searching for tools that could deliver on their promises, particularly in behavior-based threat hunting, which focuses on attacker actions rather than just their tools.
The core tension between AI’s promise and its current state can be seen across several key use cases.
| AI Use Case | Promised Benefit | Actual Concerns |
|---|---|---|
| Threat Detection | Faster response to anomalies | False positives, data overload |
| Agentic AI Integration | Autonomous security decision-making | Loss of human oversight |
| AI-Driven Analytics | Predictive defense strategies | Trustworthiness of outputs |
| AI in Phishing Defense | Real-time scam detection | Adaptability of phishing tactics |
Innovation Focuses on Identity and Trust in the AI Era
For many attendees, the RSAC Innovation Sandbox competition provided a glimpse into the future of cybersecurity. Leo Scott from DataTribe observed a clear trend among the competing startups: a sharp focus on digital identity. As AI systems become more integrated into business operations, verifying who or what is making a request is more critical than ever.
This year’s innovators prioritized solving fundamental problems at the heart of the AI trust chain.
- Many startups presented solutions for AI-related identity challenges.
- Data integrity and source verification were common themes.
- A notable shift was visible, moving from simply blocking bad actors to proactively validating legitimate users and systems.
This focus on foundational trust shows the industry is thinking deeply about the long-term security implications of AI. While major players like Cisco made headlines with their announcements, newer companies like Terra Security proved that even small, agile teams are ready to tackle these complex challenges.
Community and Public Sector Action Take Center Stage
Beyond the technology, RSAC 2025 emphasized the human element of cybersecurity. The official theme, “The Power of Community,” was felt in keynotes and informal conversations alike. As Ira Winkler of CYE wisely put it, “It’s always the same leopard. But the spots? The spots talk.” This year, the spots talked about collaboration and shared knowledge in an unpredictable landscape.
This spirit of collaboration extended to the public sector. Marcus Fowler from Darktrace Federal noted that discussions involving government agencies felt more grounded and action-oriented than in previous years. The focus was on practical issues like building workforce pipelines, securing budgets, and improving inter-agency cooperation.
Gurpreet Bhatia, the acting deputy DoD CISO, delivered a direct message in her keynote. She declared that zero-trust architecture is no longer a suggestion but a baseline requirement for federal systems. As government agencies adopt AI, the pressure to implement autonomous and fully auditable security measures will only intensify, bringing public and private sector goals into closer alignment.
Frequently Asked Questions about AI in Cybersecurity
What was the main theme of RSAC 2025?
The overwhelming theme of RSAC 2025 was artificial intelligence. Discussions covered its potential benefits for defense, the new risks it creates, and the challenge of separating practical tools from marketing hype.
What is agentic AI and why is it important?
Agentic AI refers to artificial intelligence systems that can act autonomously to perform tasks and make decisions without direct human command. It’s important because it could automate complex security responses, but it also introduces risks related to loss of human oversight and new potential attack vectors.
Are cybersecurity defenses using AI effectively against attackers?
According to experts at the conference, the industry is still in a race to catch up. Many believe attackers are currently using AI more effectively for tasks like creating convincing phishing emails and adaptable malware, while many defensive AI tools still need to prove their value.
Why was digital identity a major focus for startups at RSAC 2025?
Digital identity is crucial for securing AI systems. As AI takes on more tasks, it’s essential to verify who is making a request to prevent misuse. Startups are focusing on this area to build a foundation of trust for the AI-driven future.
