Tag Archives: research

Internet Search is Changing – Is it for the better?

Have you noticed how internet search is changing?
As long as I can remember, search worked by feeding it a few keywords or a question and it would return back thousands, or even hundreds of thousands of results. These results were prioritized by custom algorithms that take in multiple factors to prioritize them to direct you to a website for more information.
This is what most of us think of when we refer to “Googling” something. It is akin to the idea of going into a library and asking the librarian where to find a specific section of books. It didn’t lead to any direct answers, just a pointer to shelves full of books that we would then spend time searching through to find the answer we wanted.

Google Regular Search

As AI has started to take hold, I have seen how it has been embedded into our internet searches. This was a pretty cool enhancement in that the search engine would put together a simple response to your request in summary. I don’t know the statistics of it, but I am really interested to know how many searches end here, and the user never needs to click into one of the reference links.
This became more of a hybrid approach where it would give you a quick answer, but then provide a list of results like before to continue searching the book shelves for the answer you wanted.

Google AI Summary

Google’s new AI Mode takes it to the next step of a conversation. At this level, it is like you are holding a conversation with a guru of any subject. You are able to ask questions to get the answers. If the answer wasn’t quite right, you could change your question to get more specific. You keep asking questions until the response is acceptable. You will still see a list of references, related to the answer, but you won’t get hundreds of links. Think of it as if the guru gives you an answer and also provides a list of the books he used to get it. This is helpful because you may want more context or to even verify for yourself that the guru is correct.

Google AI Mode

Why do I bring this up? There was an interesting discussion on the Down the Security Rabbithole podcast, Episode 671 – It’s the End of the Internet As We Know It, that this was mentioned.
The discussion really got me thinking about how search is changing and the effects it has on the internet as a whole. I want to talk about two examples.

Effect on critical thinking
The advancement of how we search for information is absolutely amazing. The speed that we are able to find the answer to most questions has improved our ability to our jobs, grow our knowledge and help advance key research for things like technology, medicine, etc. The time it would have taken to look something up in a book was long when we were going to the library or digging through our collection of encyclopedias. I am amazed every day at the information that is available to my kids that I could have only dreamed of growing up.
I must consider the cost as well. Without guardrails in place, how do we ensure that people are still able to think critically about a problem? The results we are seeing come back from these search results can be pretty good, but they can also be pretty bad. How do people know when to believe the answer, or when to question it and dig deeper to verify the results?
The longer we go just taking the default answer from a search engine’s AI system, the further we get from spending cycles verifying the answers on our own.
These systems are getting better, but I am sure will always have their faults or build their own biases. What happens when a person asks what is 2+2 and the system returns 5? 20 years ago, I would have believed every person would question that based on their fundamental education. Today, in 2025, I believe there are people that would take this answer as fact without a question at all.
As time goes on, and people turn to Googling everything, at some point we lose that fundamental knowledge. We stop learning how to do simple addition because “why waste my time on that, when I can just ask ChatGPT?”.
Another example I give on the podcast is around cooking recipes. Have you ever just Googled how to make something simple, like chocolate chip cookies? With the new results we get from AI searches, they will just spit out a recipe. Sure, there might be some links to actual cooking sites, but why not just take the initial AI answer?
Do you know enough about food ingredients to trust the answer provided? What happens when the ingredients include something that is actually dangerous to humans? Would you know not to use that and question the recipe?
There are real world consequences to this.
Don’t get me wrong, there are similar consequences to you just following the recommendations on some random cooking site, or that influencer’s TikTok channel.

Ad Revenue
The other thing I find interesting is around ad revenue that powers both search engines and the organizations behind the links you click during historic searches. Search is free because they are making money through the ads payed by many of those links displayed and clicked. How will this affect the search providers when links are no longer displayed or they are not clicked because the AI answer was enough?
This question actually takes me back up to the pervious section where I discussed consumers just taking the AI answer and not digging deeper. What happens when the answers are biased by the target companies buying the promotions? Does this end up distorting the answers toward paid bias, and if so, how long does it take before one can easily sway the world to start believing something that just isn’t true? It just seems true because enough money was put into it to make it the theme of the answer. Then add to it that people no longer perform critical thinking and just take the answer as fact.

On the organization side, the side that pays for their ads to be in the search results, what effect do we see? If the consumer is just interacting with a bot and no links are displayed, you lose the visibility you once had. The chances of someone just finding you from a simple search decreases and less traffic may come to your site. Maybe there is some path that this tactic changes to be part of the AI results somehow, but that gets us back into the bias in those results if they are driven by sponsors.

It could just be that marketing has left search results behind. Maybe organizations have solved this through the use of social media channels. Maybe we will see commercials or ad banners start appearing in the AI chat bots.

Final Thoughts
It is an interesting time when we think about internet search offerings and how they are changing. Change isn’t always bad and the positive value in this case may outweigh the negatives. Our ability to get to answers more quickly has a lot of positive value to advancing a lot of things. However, there are some negatives that just need to be considered. We need to make sure that the responses provided are accurate and unbiased as to not start retraining our knowledge on inaccurate information. A small shift in our understandings can have a huge ripple effect years down the road.

Does “Research” Terminology Reduce Adoption Rates?

Leave a reply

What is your reaction to this tweet?

In the drive to “do something,” many applaud this as a reasonable step. I think it actually might harm our efforts and slow our progress.

Words matter.

Does the use of the term “research” reduce adoption rates vs. if we used the term QA or QC?

What is wrong with the term security research? Why might QA or QC be a better selling point?

Consider how businesses handle “research” versus quality assurance/control. In most cases, businesses have budget for quality work. They recognize the importance of producing to the level of quality expected in the marketplace.

The role of QA/QC is one of trust. Partnering together to produce a better product. A way to protect the company while growing the bottom line.

Research is a confusing concept. It either harkens back to grade school papers, college projects, or huge corporate investments. And in the corporate world, research is tightly controlled and wrought with failure. The hope is a small amount of success to make up the difference.

Research is about the future. Quality is about the current state.

Confusing the opportunity: security research

Security research is not well understood. Not even within the “research” community – Bug Bounties refer to their testers as “researchers”, “bounty hunters”, etc.Combining two expensive, confusing terms together creates additional barriers and hurdles.

Where does it fall within the budget? Is it a security item, an application item?

Does this make security testing or research bad? No. It highlights the fact that when working with an organization, perception matters.

When you approach an organization regarding security testing and approval, are they more apt to go with something that sounds familiar, they understand the value, and fits their model, or go with an option that is often interchanged with “hacker”, and they really don’t understand the value? You hear all the time how different groups need to speak the language of their consumer. While I am not a fan of the idea of all these different languages, I do think that using terminology that is familiar to the consumer provides a better connection and opportunity.

In this case, you are selling testing services. These are QA/QC services to offset the internal testing they are doing, while adding a specific focus on a limited classification of bugs. Would changing our terminology change the adoption rate?

I would love to hear others opinions on how they think choice of terminology affects adoption rate.

Hacking Cars: Taken Seriously?

Leave a reply

Turn on an ad for new vehicles and you are bound to see how connected they are to our lives. Gone are the days when your vehicle is just a stand alone product. Now we are seeing cars that have internet connectivity. We are moving past the simple satellite radio or GPS systems and becoming connected to a lot of data. Security folks have been talking about vehicle security for a while now and a few researchers have been focusing on showing how serious the security of these vehicles is.

Today, a story was released on Wired “Hackers Remotely Kill a Jeep on the Highway – With Me In It” (http://www.wired.com/2015/07/hackers-remotely-kill-jeep-highway/) describing how a Jeep was remotely controlled by a laptop 10 miles away. For the full details, check out the link I just provided. Once the story hit the airwaves, it received lots of attention, both good and bad.

Lets start with the positive side of things that were shown. It is possible to actually show the capability to breach a vehicles systems (remotely) and then control many of the functions. These functions include the radio, wipers, temperature controls, transmission and brakes to name a few. It is a concern that this can be done without authorization. I certainly do not want my vehicle to be taken over while I am driving it making it unsafe for myself or my family. The highlight: Security is important for vehicles with them being more reliant on software and internet connectivity.

Rumor is that there is a patch for the vehicle to fix this issue. The issue we now have to address is how do we efficiently and effectively get these patches to the vehicles. At this point, bringing the vehicle in to a dealership to have the software updated is the only real option.

The negative reception is where it gets interesting. They decided to do this experiment on a highway with other vehicles around traveling at the speed limit (70 MPH). At one point the driver is explaining how he can’t see because the windshield wipers are going with the fluid spraying. At another point, they cut out the transmission and the vehicle slows way down on the highway where there was no breakdown lane. That is a brief and probably insufficient summary, however the point is that a lot of people are upset.

This type of testing in a public place like this puts the other drivers on that highway at risk. This is not much different than the plane hacking bonanza that happened a few months ago (http://www.cnn.com/2015/05/17/us/fbi-hacker-flight-computer-systems/) causing a huge backlash. It is one thing to look for security issues that may help make things safer, but it is critical that the testing of these theories are done in a controlled environment, not putting people at risk. They don’t test vehicle crash ratings on the highway, they do it in a secluded area where safety is a priority.

If you are going to research security issues, no matter what they are, it is critical to think about this type of stuff before you just jump on in. While I understand that this type of stunt hacking is great for advertising an upcoming talk at your local hacker conference, it is not acceptable when directly putting other people at risk. You want to hack a plane? Get an airline to get you into a hangar in a controlled environment. The other option, by a plane to test out yourself. But don’t do it on a plane full of passengers at 30,000 feet. In this case, the researchers went out and acquired the vehicle and researched in their own facilities. The issue arose when they did their testing on a highway and not on a closed course. Security research is walking a fine line and it will require the best foot forward to push it in a positive direction. If all people see is the stunt hacking they will lose sight of the real issue at hand and just see these stunts as reckless. It will have the opposite effect of what the end goal is: to increase security awareness and security of the devices or products.

If you are in the market for a new vehicle, don’t be afraid to ask questions about the security of the vehicles communication systems. The more we dig as consumers the more aware the manufacturers will be. At some point, promoting security as a feature will be critical to beating out the competition ultimately forcing everyone to get on board. Be smart and stay safe.

Ashley Madison site breached

If you are keeping up with the news you have probably already heard about the breach of the adult site known as Ashley Madison.   Here is a link to one of the articles about it: http://money.cnn.com/2015/07/20/technology/ashley-madison-hack/index.html. Like the breach at Adult Friend Finder (http://money.cnn.com/2015/05/22/technology/adult-friendfinder-hacked/) a few months ago, this type of incident is a little different than the usual breach.   This breach is less about identity information (although at the root has a lot to do with it), financial information, or even health information.   The focus of this type of site/service is on secrecy and discreetness.   It is about sharing sensitive information about an individual’s private life.

As we become more content with the Internet and the freedoms it provides us, we often start to overlook the reality that it creates.  Some people think what they do on a computer isn’t real, rather, more of a game.   That the effects are not real.   We have this notion that everything we do is anonymous, leading us to take more risks than we might have otherwise.  Take a moment and think about some things you might have said or done online that you most likely wouldn’t have done in person.  Think about how quickly that can snowball out of control.  

The breach at Ashley Madison should serve as a reminder that what we do may not be as anonymous as we thought.   That the effects of our actions may turn out to have some real life consequences.   Is it possible someone was just curious and meant no harm signing up on the site?  Maybe they got in a fight with their spouse, had a bad day at work, were just bored.   Of course those may not be acceptable excuses for joining a site that promotes adultery, but it could be something that small that led to the initial curiosity.  There are also people just looking for another relationship.  Anyone who has their name released as being a member has the same potential consequences.  You may be publicly criticized, sorry.. that is what society does now.  Your job or career could be effected. Your relationship with your significant other and/or children can be effected.  The list goes on.

We are all still learning the effects our online actions have on us over time. Our parents didn’t have nearly as much technology so many of us are learning on our own. We need to understand that, just like business, we assume a level of risk when acting online.  There is no 100% secure systems.  It doesn’t matter if we are talking online banking, adult sites, social media, or password managers.  There is always some level of risk.  We must learn to calculate that risk and determine if it is worth it.  We are often quick to blindly accept the risk for the quick reward.  Share your contacts for some coins in a game.  Post atrocious comments for a chance to feel like you stood for something.  

Is there a risk to joining an adult site like this?  Of course there is.  For many, that risk is acceptable for their own personal reasons.  Some members may had not really considered the risks, while others may have given great thought to joining.  Either way, the risk is there.   The big question in this situation is regarding what that risk now means to the individuals involved.

The media hypes this up to be devastating.  However if we look back at Adult Friend Finder, after a few days, you stopped hearing much about it.  This doesn’t mean that there were no consequences suffered by users effected by that breach, but it did quiet down a lot.  Maybe it was because of the personal nature that people didn’t want to put it out for everyone to see.  That makes it difficult to judge the real effect that this breach will have.

It will be interesting to see what types of effects this has going forward.  In the meantime, we should ensure that we are thinking about the risks. Be safe everyone.