Google Searches and ML-Could Google Become Largest Inside Trading Platform Ever?
Using the latest technology, overnight Google can position itself to leverage big data to perform insider trading against public companies on an unprecedented scale - with no government oversight.
Whether you are a publicly traded company or have a modest 401k, you need to start asking how to protect yourself and your organization.
Today, everyone seems concerned with individual privacy as it relates to the information that Google and Facebook is retaining on each of us. Tech-savvy individuals with time on their hands can download the Google Takeout file associated with their Google Account, which shows each and every search executed over the past several years. Since Google’s existence is based primarily upon data, it is intelligent enough when an individual leverages multiple platforms (laptops, desktops, mobile), to consolidate all of your searches into a single repository.
Unless you’re living under a rock, you realize that with Google you are the product and that they sell you and your information to advertisers.
Let’s take a minute and think about what Google is collecting from the searches I run. Like many of us, I have a work laptop and an Android compatible device. On my laptop, I use Chrome because I prefer it over Internet Explorer and I also have a gMail account for personal reasons (or even a gMail business account) linked to my laptop.
Each and every day, as I go about my activities, I use Google to perform most of my internet searches. After all, who can live today without a search engine? I cannot recall the last time I looked at a card index system in a library or searched for something using the Dewey Decimal System.
For the beneficial and free service it provides me, Google in return tracks what I am requesting with each search. This morning I came into the office and perform the following searches using Google.
Device Date/Time Search
Laptop Mon 8:00 AM Whitehouse Correspondence Dinner
Laptop Mon 8:43 AM MBA Secondary Conference 2018
Laptop Mon 8:51 AM Marriott Marquis New York
Laptop Mon 9:12 AM Best Steak Houses in New York
Laptop Mon 9:23 AM Gallagher’s Steakhouse
Mobile Mon 11:23 AM Pizza Delivery Near Me
Mobile Mon 11:56 AM Cheap Broadway Show Tickets
Each and every day I continue with the same typical pattern. Go to work, perform my job function, and leverage Google for almost all searching required, both personal and business. Days turn into weeks, weeks turn into months, and months turn into years. That is a LOT OF SEARCHES!
There is one more critical piece of information that google is tracking and that is where my searches or ‘traffic’ originated. To be fair, it needs this information to be able to return the search results back to me. If you are not aware, Google states it tracks the following information on you (things you search for, websites you visit, videos you watch, ads you click, your physical location, device information, IP address and cookie data).
Google continues to collect this information on me and builds a profile of Norman Gottschalk. Before too long, Google knows that I prefer Wendy’s over McDonalds, my favorite color is orange, I spend most of my time in Pittsburgh, and that I work 70 hours a week because it knows where I am logged into a computer.
Wait – how does Google know I work 70 hours a week? Oh- that’s pretty simple for them since I am logged into Chrome at work and they can interpolate that by the time of day, the device, content of searches and the frequency that I perform searches.
So now things are getting interesting – why? Because I just intertwined the personal privacy with my work. Google knows by my search history when I am at work and what I am working on. And if Google doesn’t know it is me specifically, it at least knows that it is someone at my office.
Remember earlier, I stated that Google knows where the search originated from (if you doubt this, check your Google takeout search history takeout.google.com) and find out for yourself.
At this point, we all should realize that Google tracks my searches and has built a profile on me with very little consent from me. But after all, I guess I can’t argue since the searches are FREE.
Now for the bigger picture.
If Google has built a profile of me, including my physical search locations, then it is reasonable to assume to Google has also built a profile on my coworkers. But how does Google know who my coworkers are if I haven’t shared my work email with Google?
Let’s pretend I have no social medial profile whatsoever (in reality, I only have a LinkedIn account) - so again – how does Google know who I work for? This can often times be answered by where my internet searches originate from – you know that pesky term IT resources like to use IP Address. Every IP address most likely will be registered to a physical location and will include reverse lookup records to help identity what company is behind that address. If your specific IP address doesn’t have a reverse lookup record there are numerous other public sources to identify your organization.
Now, let’s turn our thoughts away from thinking about my individual searches, and let’s pivot the data Google has against the physical location of a search. In other words, instead of looking at my profile let’s look at the search history profile of the company I work for based upon the address where my traffic originates.
Collecting all the search request data from my company over a period of time may provide valuable insights into what is happening within the organization. This information and the potential search trends within the company may just provide financial insights that could be considered confidential.
Now, imagine I am the CEO of the publicly traded ACME Horseshoe Company out of East Brady, Penn. ACME has $75 million in annual revenue, operates 300 stores predominately in rural markets and employs 1,900 employees. Unfortunately, there isn’t a large market for horseshoes any longer, and as the CEO of the company, I realize that we are quickly entering into significant financial hardship that will force the organization to quickly reduce staffing, close stores, cut expenses, etc. As the CEO, I then hold a leadership meeting and explain to the executives that we are going to have to significantly cut expenses, close stores and layoff individuals. In an effort to avoid bankruptcy, I start seeking new lines of business to enter that will have synergies with my company.
After the leadership meeting, my executive team leaves and each one begins to perform the relevant tasks that they were assigned. Some research cost cutting, some research store closings, while others begin to research bankruptcy and WARN ACT, etc. Before you know it, the Google Search Profile on my company begins to appear as follows:
ACME Horseshoe Corp
Viewing one individual executive search information, may not paint the clear picture of the organization. However, viewing the searches from 10 different executives within the organization, it is clear to see that the company is most likely having financial problems.
Using historical search terms of companies that have filed bankruptcy within the retail market, I can use current Machine Learning technology to identify similar search patterns among these companies. Before too long and with relative ease, I can build up an Artificial Intelligence “AI” Support Vector Machine “SVM” to identify the probability of a company entering into financial hardship over the next 60-90 days.
Armed with this new Vector Machine I have created, and executing this AI approach against the search results for an individual IP address (which in essence relates directly to an organization), I may now be able to identify when a company will have mass layoffs, file bankruptcy, or have other significant events that could affect the stock price. Once I have identified this event, then I can easily short the stock on the market and simply wait for the announcement.
Being the greedy individual I am, I decide that since I can do this so easily I am going to monitor the top 100 stocks daily. Estimating it will take me 30 minutes to perform the necessary research to identify the target IP addresses, once that is done my Machine Learning Solution will daily send me negative sentiment report of the stocks I am monitoring.
Here is one of the most important parts of this. I am not basing this decision on what the public company has officially announced, I am basing this information on the actions of the employees within that company, potentially long before the market is made aware via a public announcement.
As easily as I have described a negative event begin tracked and reported against a company, Machine Learning and Artificial Intelligence can also identify positive events that would allow an individual or organization to financially benefit. All that is needed is historical data to build the base model, and a daily feed of search engine requests. Again, keep in mind, I said Search Engine Requests, not Search Engine Results, or click troughs. Click troughs will only help to validate the searches.
While today Google doesn’t provide these search term requests based upon source IP address that I am aware of, there is nothing preventing it from happening tomorrow or in the future. Regardless if Google sells this data today, elevated or privileges resources inside of Google most likely already have the ability to execute these exact queries today!
Who knows - maybe some Google employees today are simply getting wealthy playing the market with this type of insider information (after all, the search originated inside of the organization). This would be done very simply by questioning a overlapping list of companies who searched for “WARN ACT” and “BANKRUPTCY” from the same IP address within a 7 day period, and then identifying where the request came from.
Taking this even further, what if Google decided to advance its personal agenda against companies using this type information? Seem too farfetched? I personally don’t think so in today’s day and age. We are all aware of the accusations against Google having a bias towards Hillary Clinton in the 2016 Presidential Elections. So - Elections today, tomorrow it could be Gun Manufacturers, cigarette manufacturers, or even manufacturers of stuffed animals.
I am the last person in the world who would suggest government regulation – however, with this amount of data at their fingertips, Google, its executives and engineers could easily take control of the stock market on what one would almost call sure bets!
I am fortunate enough to work for an organization with a robust innovative culture and that prides itself in out of the box thinking and understanding the big picture on how organizations grow and protect themselves. It has been through this environment - with our competencies around machine learning and security - that I have come to this realization. Now that I’ve shared it with you, how will you respond? What will your organization do differently? Not sure? Just reach out to me.