As the U.S. news media drones on about Russian collusion within the U.S. political class well into a second year, it is becoming increasingly clear that the intelligence product that lit the spark of this fire – the now infamous dossier created by British ex-intelligence officer Christopher Steele – is more a work of politics than good intelligence. Just as in the sports and entertainment industries, politics ruins the business of intelligence.
While intelligence is used extensively to understand political threats, using political ideology in the creation of intelligence is a scourge on the business. Politically charged products, like the Steele dossier, may look like a valid and well-constructed intelligence report on the surface, but eventually fall apart when placed under the scrutiny of other intelligence information.
The underlying difference between good intelligence and politically tainted intelligence lies in the purpose of their creation. Intelligence exists for the use of a customer – someone who needs intelligence about a threat or an adversary in order to make informed decisions. For intelligence support to be effective, it must inform customers in the most unbiased manner possible so they can make their own decisions about a given problem set. Intelligence assets are intended to support a customer – a decision-maker – by using sound and proven collection, along with strong analysis and critical-thinking skills.
While intelligence seeks to inform decision-makers, politics seeks to influence them. A politicized product attempts to get a customer to agree with a decision or perspective reached by the creator within their own minds. Politically biased analysis does not equal truth and accuracy. Instead, analytic support containing political bias taints sound analysis with a layer or filter of ideological principles or motivations. Because these layers are mostly transparent, spotting a political report amongst the daily flow of intelligence is extremely difficult.
Risking Customer Trust
Whether intentional or not, intelligence products that include politically biased information in their analysis and assessments run the risk of being inaccurate, vulnerable to deception, or intellectually lazy. Those who include or allow it repeatedly in their work risk being perceived as deceitful or careless by their customers, ultimately losing trust and confidence.
This trust is very hard to attain and even harder to regain if lost. If a customer believes they are getting analysis that was guided by spin and ideology instead of sound methodology, they will soon turn to other sources to help them make decisions. The only persuasion an intelligence professional need worry about is to persuade their customer that their work is credible, accurate, timely, and without bias of any sort.
Identify, Mitigate, and Minimize the Spin
A vast majority of intelligence analysts don’t intentionally add political bias into their own analysis, but that doesn’t mean that their intelligence products can’t be used to further someone’s political agenda. There is little that an intelligence professional can do if their supported decision-maker decides to make a political argument with their product. What’s important for intelligence analysts in this era of fake news, spin, propaganda, and outright deception is to take active measures to ensure their own work isn’t tainted by theirs, or someone else’s, political ideology.
While there are many methods and techniques to help the individual with their own analysis, there are several operational steps that managers often overlook.
1. Identify Where it is Most Likely to Be Found
At the tactical or street level, politicization of intelligence is extremely rare. Spot reports, moving target indicators, tactical intelligence estimates, network diagrams, and threat maps are simply tools for analysis in support of operations. These are raw pieces of data that have not been pre-processed or edited so they haven’t been touched by human bias and, as such, are not suited to directly or indirectly pose a political argument.
The higher the level of intelligence support, to include the combatant command and national-level agencies, the more effort there should be within those organizations to ensure that political bias is identified and exposed. As the Trump dossier has highlighted, any intelligence containing information that is already under intense political debate is suspect. Intelligence organizations that rely on information that has been collected, processed, handled, and packaged with someone else’s analysis – especially open-sourced analysis – should also be heavily scrutinized. If much of the information analysts rely upon is second-hand analysis, time and effort should be made to evaluate these sources. As recent news has shown, both commercial and government sources can be politically biased.
2. Minimize Personnel and Organizational Bias
Notice that I did not say to eliminate your own bias; that’s impossible. We all are equipped to make discriminating judgments in our lives, and many of those require us to take our emotions into account. Unless you’re a Vulcan, it is impossible to totally eradicate emotional perspectives from your analytic thought processes. Instead, recognize your own perspective or have someone else tell you what they see as a pattern in your opinions and ideas. You may not like or agree with what they see, but it is still crucial to periodically get outside of the echo chamber of your own opinions. With this knowledge, you can better identify any of your own bias that may happen when creating your own reports or products and adjust to minimize or better explain it as you work.
It is also critical for intelligence managers, team leaders, and first-line supervisors to identify and minimize bias in their own people. One of the most effective tools in this regard is to build in diversity of backgrounds, opinions, and perspectives among employees: A wide representation of ages, sex, cultures, ranks, education levels, ethnicities, religions, among a team of people may create a challenging work environment, but will also help to mitigate myopic or bandwagon perspectives.
3. Require Bias Mitigation Steps in Analytic Processes
While there are many structured analytic and critical thinking techniques to help the individual analyst review their sources and quality of work, these are often overlooked or simply not used because of either the arrogance of the analyst or the pace of their operations. Just like committing to get in shape physically, an intelligence team, section, or organization has to build in time to make their analysis healthy and trim the political fat from their products.
Larger organizations can rely on production and quality assurance managers, but every organization must incorporate required steps or processes to reduce bias. These steps could include everything from simple vetting from another analyst to the use of contrarian techniques that require collaboration and checks on assumptions, hypotheses, or competing ideas.
4. Train Your Analysts and Your Machines
Merely training analysts on the “buttonology” of data mining and visualization tools is a near guarantee that your analysts will not be equipped to identify and mitigate bias in their products. Training on critical thinking and structured analytic techniques gives them more tools in their analytic tool box to draw from and apply in the different situations they encounter. If your analysts have access to a training program, take a look at from the manager’s perspective – does that program train analysts on how the engine works, or does it train them on how to drive the car? If it’s the former, insist or look for more applied learning for your people.
In addition, review the search engines or entity extractors that your analysts regularly use. Is there a bias in their search algorithms? I once used a tool that used a proximity algorithm to automatically link two subjects if they were five words or less from each other in the same sentence. This caused an incredible amount of false positives and simply erroneous associations that took too much time to weed through. As recent revelations from Facebook and other social media sites have exposed, the tools that search and parse through big data sets can contain both technical and political bias. They were written by humans, after all.
Paraphrasing an old adage used when I was in the military – opinions are like belly buttons (or something similar), everyone has them. For intelligence, any piece of intelligence-related information that is used that has already been worked on by another human can contain political bias in any number of variations. While you cannot do much to prevent your intelligence being used for political purposes, it’s important that analysts and managers take necessary steps to identify and minimize it within their own work.