In this forward-looking, tell-all interview, Mazhar Hamayun, a Check Point Regional Architect, provides insight into the profound concerns surrounding the rapid growth of AI, delves into how we can effectively address these concerns on multiple different levels, and offers a fresh lens through which to interpret exciting, AI-related innovation. Explore thought-provoking perspectives that can enrich your endeavors, help you safeguard systems and enable business growth.

Why are people worried about the rapid development of AI-based technologies?

The acceleration of AI’s development has elicited both awe and concern in equal measure. Among the prominent reasons for the latter is the anxiety about potential job displacement.

As AI systems become increasingly sophisticated and capable of performing complex tasks that were once reserved for humans, there’s an apprehension that a significant portion of the workforce could be rendered obsolete, sparking widespread unemployment.

Moreover, as AI interfaces like chatbots begin to exhibit human-like intelligence, questions arise about our role and significance in a world increasingly dominated by machines.

The fear isn’t just that these systems might surpass us in specific domains of intelligence, but rather there’s an existential dread that they could eventually eclipse us completely, resulting in a displacement of humanity from its unique position in the existential hierarchy.

How might lack of transparency and explainability in some AI algorithms fuel worries about use of AI for decision-making (cyber security-related and otherwise)?

Transparency not only allow us to understand, predict, and correct the behavior of AI systems, but also to establish trust, uphold legal and ethical standards, and ensure security. AI algorithms can certainly fuel worries about their use for decision-making. Here’s an overview of how:

  • Bias and discrimination: Algorithms are trained using data that could potentially have biases, which, when not controlled, can lead to discrimination. For instance, if an AI system used for hiring processes was trained on data that contained gender bias, it could potentially discriminate against applicants of a certain gender.
  • Unintended consequences: Without transparency, it’s difficult to predict how an AI algorithm will behave in all situations. For instance, in cyber security, an AI algorithm might be designed to detect threats, but without a clear understanding of its decision-making process, it could flag harmless activities as malicious, leading to false positives.
  • Accountability: If decisions made by AI go wrong based on badly trained algorithms, it can be difficult to hold anyone accountable. For example, if an AI system involved in autonomous driving makes an incorrect decision that leads to an accident, lack of transparency in the algorithm’s decision-making process could make it hard to determine responsibility.
  • Trust: Trust is the biggest factor in the use of AI technologies. If users do not understand how decisions are being made, they may not trust the technology. In fields like cyber security, where decisions can have substantial impact, this lack of trust can be a significant obstacle.
  • Legal and ethical issues: Lack of transparency can also pose significant legal and ethical challenges. Laws like GDPR require certain levels of detail and transparency from AI systems. If these requirements aren’t met, organizations could face legal penalties.

How can CISOs and cyber security professionals learn about the data used to build their AI tools?

CISOs and cyber security professionals must have a thorough understanding of the data’s origin, type, and the applications involved in the building of their AI tools. This insight is indispensable in strengthening cyber security measures, eliminating bias, and fostering fairness. Here are few quick ways to find out about data used to build AI tools:

In-depth review of vendor documentation: Go through the comprehensive reports and documentation provided by AI vendors. They should give detailed information about the data’s source, type, and how it was processed. This review will be instrumental in understanding the AI system’s training process, the bias mitigation measures in place, and other significant aspects of the data.

Opt for third-party audits: Seek out independent audits of the AI system. These audits offer an impartial perspective on the AI system’s data, algorithms, and overall performance, giving a more transparent view.

Implement PoC and conduct testing: Run proof-of-concept or pilot tests of the AI tools within a controlled environment. These tests offer invaluable insights into the tools’ operation and can shed light on the nature of the training data used, even if indirectly.

What should CISOs and cyber security leaders tell their higher-ups about transparency of AI tools, in your view, as this relates to the overall ethicality of the business?

CISOs and cyber security leaders need to assertively champion the cause of comprehensive transparency in AI tools. We must consider that AI’s opacity, specifically concerning data sourcing, processing paradigms, and decision-making algorithms, can be a profound ethical concern and even a potential operational risk.

When we establish full transparency, we are building a stronger foundation of trust among our crucial stakeholders, clients, and regulatory bodies. But this isn’t solely about gaining trust; it’s also a practical necessity to ensure adherence to a progressively stringent legal framework surrounding AI and data use.

Furthermore, adopting a proactive stance on transparency can enable a more effective risk management strategy. A clear understanding of AI functionalities can assist in identifying and addressing potential vulnerabilities, ultimately fortifying our cyber security infrastructure against future threats.

How does the potential invasion of privacy through increased data collection (and surveillance) via AI systems worsen fears around AI?

The enhanced capability of AI systems to collect and analyze massive amounts of data presents both opportunities and challenges. One significant concern that arises from this capability is the potential encroachment on individual privacy, which can amplify existing anxieties about AI.

AI’s ability to extract and infer from an extraordinary volume of personal data, particularly in instances without clear consent or awareness, can give rise to apprehensions about unauthorized access or misuse. Such fears are heightened by the absence of transparency around how AI processes and makes decisions based on this data.

In essence, the balancing act between the value proposition of AI and the preservation of privacy is critical. It is our responsibility to ensure that this balance is struck, fostering trust among our stakeholders, and mitigating fears surrounding AI’s implications for privacy.

How are producers of AI tools and those who deploy AI tools addressing concerns around privacy, if at all? 

The strategies employed by AI tool developers in order to manage privacy concerns are advancing. Key among these strategies is the adoption of Privacy-by-Design principles; a proactive measure that ensures that safeguards are embedded within AI products from the inception, rather than as an afterthought.

In addition, efforts to improve transparency and to articulate clear, easily understood privacy policies represent steps toward ensuring informed user consent and fostering trust. The commitment to regular audits and ethical reviews serves as a necessary check to ensure adherence to evolving privacy standards and laws.

How should policy leaders address these concerns, in your opinion?

Policy leaders carry substantial responsibility when it comes to mitigating the growing concerns arising from AI’s increased use. Addressing concerns necessitates an effective strategy, at the heart of which should be the formulation of robust, yet adaptable policies.

These policies need to address pressing issues, such as privacy and data protection, while also actively deterring AI bias. They should be comprehensive, but retain the flexibility to evolve in step with AI advancements, thus eventually fostering innovation without compromising ethical standards. A strong oversight mechanism, accompanied by regular audits, is essential to ensure that AI systems comply with these policies. This not only protects user interests, but also maintains the integrity of AI operations.

Is there anything else that you would like to share with the Cyber Talk audience?

AI and its utilization in daily life is not just about technology. It’s about expanding creativity and human potential. Let’s use it to solve complex problems, foster innovation, and create a future where technology works to everyone’s benefit.

For more insights from Mazhar Hamayun, please see’s past coverage. Lastly, to receive more timely cyber security news, insights and cutting-edge analyses, please sign up for the newsletter.