Security Questions in the AI age

I seem to be attending quite a few security/AI/platform engineering events around my place and honestly, it has been a delightful experience. For a person who is always writing remotely on security/technology topics, meeting people and sharing ideas is a refreshing thing. There are many things that cannot be learnt from an online environment alone. An offline environment and sharing ideas and understanding different viewpoints is great for learning and is fun.

These were some of the questions that I encountered during some of these talks:

1. In the AI age, is it worth studying traditional cyber security practices? (after all, AI has the capability to do a lot of security activities even more easily, right)
   • Just like other fields that have been automated or simplified by AI, cybersecurity has also faced automations and simplifications with AI. But it is not to say, that everyone and everything will automatically move to AI now. It is a gradual process, and one has to understand traditional security concepts, tools and practices to really understand what AI is doing and if it is really doing it right. So, students will have to learn basic cybersecurity concepts like defense in depth, firewalls, anti-virus definitions, pattern recognition, OAuth, authentication and many other fundamental concepts. This will enable them to understand AI security better.
2. How can one get cyber security entry level roles when not many companies come for such roles in colleges?
   1. It is true that college placements do not feature a lot of companies who come for entry level cybersecurity positions. For passionate cyber security students, who only seek a cybersecurity career, it is good to attend hackathons and cybersecurity/technical meetups where startups might be present. Startups will be open to hire fresh college students. The salary might be less, but the experience gained through the job is extremely valuable. Having connections on LinkedIn who might post about entry level cybersecurity positions is another way to get one’s foot into the cybersecurity door. Having a curious mind, observing, studying and tinkering safely are some ways to find entry level cybersecurity positions in the college level.It is also good to try and get some entry level cybersecurity certifications that might bolster one’s resume even more.
3. Who is responsible if AI makes mistakes?
   1. ChatGPT broke into the world only about 4 years ago and we have slowly learnt the nuances of working with it and other AI models. It has been helpful for many of us, and it has been a burden for others. AI has been seen to make medical mistakes, hallucinations, producing wrong results, judging incorrectly and many other mistakes.
   2. After all the mistakes that AI has made, it is important to discuss accountability as well. If AI makes mistakes, it cannot be pinned on it alone, since it is just a machine acting on somebody else’s directions.
   3. So, as of today if AI makes mistakes, it is a joint responsibility between the developers, the data scientists, the users (for asking the questions and accepting the answers in blind faith) and the organizations who deployed the employers in the first place.

These were some questions that were asked in general and these are some of my answers. Do feel to offer your insights as well and hope to discuss some more questions in the next few posts.

This post is part of Blogchatter Half Marathon