AI and Cybersecurity: Navigating Promise and Pitfalls

We recently held a webinar to discuss the promise and pitfalls of using AI for cybersecurity. We had a knowledgeable panel consisting of CJ Hurd, the SVP of vCISO services at Compliance IT, Chris Willis, the VP of cybersecurity and network engineering at Sayers, Linn Freedman, chair of the data privacy and cybersecurity team at Robinson+Cole, and Katie Truchon, assistant professor in the IT department at the New England Institute of Technology. During the webinar, we covered many important topics around AI and security, including the impact on jobs, legal considerations, adversary tactics, ethics, and more.

AI’s Impact on the Job Market

A common question around AI is how it may affect jobs, especially in cybersecurity. The panelists agreed that human oversight remains critical while AI can help automate certain mundane tasks. As Chris noted, AI allows us to focus more on strategic initiatives rather than repetitive manual work.

However, as CJ pointed out, AI may reduce the need for some entry-level security roles over time. The increased automation from AI tools could supplement some of these professionals’ basic work.

That said, the panelists emphasized there would be an elevated demand for oversight positions higher up in organizations. As AI takes on more basic responsibilities, humans will be freed up to review, audit, and ensure the AI systems are working properly. So while certain jobs may change with AI, new opportunities will arise around governing these tools.

Proper training remains essential too. As Katie explained, students hoping to leverage AI as a shortcut still require foundational knowledge of areas like programming. Understanding how the AI recommendations fit together technically continues to be a human task.

Applicable Laws and Regulations

Given how new the widespread use of AI is, the regulatory environment remains uncertain. As Linn explained, no national AI regulations exist in the United States. However, organizations like the White House, the NIST governing body, and others have released guidance around responsible development and adoption of AI.

While broad federal regulations may take time, we can expect to see more state-level policies emerge in the shorter term. States often pioneer consumer protection laws, and AI oversight could follow a similar pattern. The panelists noted that AI requires specialized governance that generalized privacy rules don’t fully cover.

As Linn summed up, transparency and assessing vendor risk are crucial when deploying these tools responsibly. And though clear-cut statutes don’t yet exist, staying abreast of guidance as it evolves is key.

How AI Can be Used by Attackers

Unfortunately, just as AI can provide defensive advantages, attackers leverage it to further their goals too. As Chris explained, capabilities like stronger social engineering campaigns, faster data analytics, and better evasion of protective measures are now at criminals’ fingertips.

CJ also highlighted the explosion of ransomware over the last few years. The ease of deploying more sophisticated attacks with AI may be one driver of this trend. With higher likelihood and rising impact, the risk matrix for organizations is changing quickly. This is especially true for smaller businesses with more limited security resources.

The panelists clarified that AI-powered threats pose a serious challenge to cybersecurity teams. As AI proliferates, we’ll have to work diligently to counteract these concerning new attack capabilities.

Ethical Considerations for AI Monitoring

AI has many beneficial uses, but certain applications warrant extra caution. When discussing options like surveillance and monitoring, we have to weigh risks to individual rights and liberties carefully.

As CJ noted, expectations of privacy have faded in many online spaces with the proliferation of user agreements and system notifications. However, as Linn highlighted, state-level statutes still regulate employee monitoring. Consumer protections for personal data also continue to expand. So while transparency remains important, observing both ethics and evolving laws together is critical.

Though AI surveillance may offer more advanced security, keeping people’s best interests in mind has to remain a priority. Responsible governance requires updated guidance tailored to these innovative tools.

Leveraging AI for Incident Response

While adversaries utilize AI to achieve their goals, AI can also help cybersecurity teams respond to incidents more efficiently. As CJ explained, threat detection and analysis are some of the biggest areas for improvement with AI. Quickly identifying, understanding, and isolating threats depends heavily on effective data processing.

Chris expanded further, noting that gathering, correlating, and making sense of all the relevant event logs and indicators of compromise becomes far easier with AI assistance. Determining scope, impact, and optimal response steps leverages these platforms’ data parsing strengths.

With the average dwell time of attackers hovering around 200 days by some estimates, rapidly decreasing incident investigation and containment windows is hugely impactful. AI promises to help organizations cut response times from months down to minutes in some cases. This game-changing speedup makes AI a valuable tool for any modern SOC.

Reliability of AI for Finding Vulnerabilities

Considering how effectively malicious programs evade traditional signature-based defenses, AI offers better means of identifying novel threats. As Chris explained, top cybersecurity products now lean heavily on machine learning and artificial intelligence to track new attack patterns. AI-augmented scans catch in-the-wild threats traditional systems miss by detecting harmful behaviors rather than specific code snippets.

However, Katie cautioned that AI cyber tools still demand thorough vetting and testing. Out-of-the-box solutions rarely suit every organization’s needs. While AI systems have achieved major improvements in threat detection, ensuring any new solution integrates well and provides sufficient visibility remains essential. Understanding an AI tool’s strengths and weaknesses requires deliberate evaluation by knowledgeable security personnel.

Training the Next Generation

Finally, the panel discussed how educating professionals to develop, manage and regulate AI systems remains an emerging need. As Katie explained, many colleges are adding AI fundamentals and security courses. Technical build-it-yourself experience creates better foundations than merely using AI as a black box.

Linn also noted the necessity of legal training around privacy and liability. Law programs are only starting to offer specializations addressing AI specifically. For now, law firms often handle AI via ongoing education of interested attorneys. However, formalized curricula and professional concentrations will grow to meet demand in the coming years.

With AI set to impact so much of cybersecurity and technology regulation, having multidisciplinary expertise across these areas is crucial for the next generation of practitioners. Students pursuing this space must complement technical skills with policy and compliance knowledge. Organizations implementing AI likewise need integrated teams with legal, ethical, and security experience collaborating closely.

With knowledge of the applicable developing laws and adversary tactics, we can leverage AI’s immense potential while responsibly mitigating emerging risks. Combining compassion and understanding of this technology’s implications for people’s rights and welfare, we can move towards more secure and equitable AI adoption.