Are there security downsides to the newest Technologies?

As we close the chapter 2019, I would like to share some of the things running through my mind as regards the trending technologies and security. Recently, we were invited to a birthday party at Lekki. On the fixed date, as usual, I launched my google maps app and searched for the venue. Immediately, I got an estimated arrival time, best route, traffic situation etc to the venue. I was glad, it’s a Sunday, traffic was good quite unusual for a busy Lagos and confidently told my kids, we will be at the party in 54mins. We drove, obediently following Google maps directives until we got to the address. Sadly, it wasn’t the location. Apart from being confused that we had lost our way, I had to quickly think of the best explanation to give to the 3 party-anxious kids at the rear seats.

Now, my personal experience explains one of the challenges with trending technologies. Trending technologies are expected to reign at least for the next decade. Technologies such as Artificial Intelligence, Internet of Things (IoT), Intelligent Apps (I–Apps), Big Data, RPA (Robotic Process Automation) etc. Technologies believed by many that they pose one of the greatest threats to humans while others believe that they are the best thing to happen to mankind because of the numerous opportunities. Opportunities such as prediction in business patterns and weather forecasts, even in business operational processes automation and even medical treatment and pharmaceuticals.

Artificial intelligence as a technology is any system that exhibits behaviour that could be interpreted as human intelligence (Artificial Intelligence Foundations, Doug Rose). The concept of AI especially has seen in science fiction movies has led a lot of us thinking that Robots i.e. artificial humans will soon erase humanity. However, while this is unbelievably untrue, there are some security issues that have emerged due to the new technologies such as AI.

From my personal experience above, we ended up in a wrong location because of the incorrect data provided to the app, a missing name in the address shared. Apparently, the actual location of the address had a third name. That is, Alice Johnson Street was the address that was put into google map instead of Alice Wonder Johnson Street which is the actual address. Coincidentally, there is another Alice Johnson street. The rule for using Google map to get to the location correctly had failed! AI is based on rules and pattern-matching.

AI simply flexes its special talent of following rules and pattern matching. The security implications are not farfetched. For example, AI is widely used in the healthcare industry, we're beginning to see surgical robots and other robots that provide medical supplies from one part of a hospital to another. What if the wrong medication was accidentally inputted? Or in a case where based on the symptoms entered into the system, the machine suggests another diagnosis and treatment, the implication can be disastrous to the patient. This explains the very first security challenge I would like to talk about of AI.

Security issues highlighted in this article are:

·      Inaccurate Setup and Usage: The fact that AI is rule and pattern-based means that it will work perfectly according to the data fed into it. Imagine a bot designed to update a web application from a database. If it is set or linked up with the wrong database, it will update the form with the details. Wrong setup and usage means wrong output from AI. Robots or technology at large however smart or intelligent will work based on rules and historical data. This was the case of my experience above.

·      Inappropriate design and configuration: This is closely related to the first issue raised. Even if a bot is fed with the appropriate and correct data. A badly configured or designed machine will not work according to the required specification.

·      Increased landscape for cyberattacks to thrive: Now, many security teams in organisations are working tirelessly to ensure their current systems are secured. i.e. vulnerabilities are closed, patches are updated, different threat models are set up for different systems to determine the associated cyber risks. A lot of efforts are put in place to ensure that systems are safe. With the advent and continuous evolution of new technologies such as bots, big data and AIs, there are more attack surfaces at risk and as a result, more systems to be maintained.

·      Cyberattacks: Just like every other system, an AI system or BOT can be attacked. Imagine a MiM attack or sabotage on a surgical robot just before a surgical operation. What if the information required for a successful operation was changed into something else? The effect will be grievous. You could also consider if a bot designed to update the balance sheet of company was attacked to automatically add an additional zero to each amount in the balance sheet. The effects could only be better imagined to a financial organisation.

·      Finally, the issue of privacy: With AI, personal information is shared with more systems that ever before. From Siri the AI on most handheld devices and Alexa, a virtual assistant, which are undoubtfully very useful, to YouTube and Netflix knowing our kind of music and movies. More personal information is shared with more systems than expected either knowingly or unknowingly. A real-life privacy incident happened when Netflix published their dataset. While the data was anonymized, hackers were able to identify the authors of reviews (AI Security and Adversarial Machine Learning 101, Alexander Polyakov).

Well to end my story, a call was put through to the celebrant and we confirmed that we were at the wrong location. The actual GPS location was shared with us via Google maps 😊. To say that the latest technologies are undoubtfully very useful if designed, configured, used and maintained appropriately is an understatement. Security must always be considered in all emerging and trending Technologies. Thanks for reading and have a prosperous new year in advance.