Resources

Technological Tools and Resources

Many technological tools have the capacity to impede on individuals’ personal privacy and data ownership rights. Tools can also contain bias in their operations, whether intentional or not. Please ensure that you address any ethical concerns that your solution could face. It is important that your solution respects individuals’ privacy and is fair, accountable, and transparent in its approach.

Automation

Physical Control Automation

The use of various control systems for executing tasks such as operating equipment machinery, processes in factories, switching on telephone networks, steering and stabilization of ships, aircraft and other applications and vehicles with minimal or reduced human involvement.

Chatbots

Allows open-ended or guided dialogs with subjects, over many possible channels including SMS, Skype, web, IVR and others. Uses natural language processing (NLP) to interpret spoken or written input, and respond with conversational output.

Simulation

Simulation is a useful tool for training as it allows people to engage and learn by doing. Simulations can also be built so that the situations differ enough to present new scenarios and tasks, forcing the trainee to continually think and adapt to the simulation.

Scripting Systems

A programming approach that automates the execution of tasks that would alternatively be executed one-by-one by a human operator.

Data Collection & Communication

Smartphone sensors

Modern smartphones contain many sensors (audio, light, GPS, orientation, motion, temperature, etc.) and these can be scripted to interact with apps and remote APIs when conditions are met. On-phone scripting provides a low-cost means to report on subject’s activity and provide interactivity.

Digital Activity Tracking (social media, email, and online activity)

With more and more of our activities occuring online, systems have the ability to gather and utilized incredible amounts of data on user activity and preferences. This vast amount of information can feed artificial intelligence and machine learning systems to for a variety of use cases.

Physical Activity Monitoring

Improvements in sensors and reductions in their cost means that it is now possible to incorporate them into clothing and wearables to gather data on walking, running, and many other kinds of motion and physical activity. Sensors can also monitor physiological systems like heart rate and algorithms can be trained to utilize this continual flow of information.

Geo-Fencing

Provides ability to create custom zones on a virtual map, and to then track people’s actual entry into, departure from, and duration within that zone. This functionality is generally coupled with smartphone sensors as it requires a person to have a GPS-enabled phone and app to report position their position.

Bluetooth Beacons

Small devices that emit a low-power, short-range Bluetooth signal that can be detected by smartphone apps or other Bluetooth enabled devices. Placed throughout a physical environment they behave like tiny lighthouses; that is, passing smartphones detect and report their signals to a central server, which communicates the smartphone’s location..

Data Processing

Image and Facial Analysis

Provides analysis of photos, faces, drawings and videos for items like the visual content (e.g. the physical items visually depicted), the number of people, the age and gender of people, and the activities of people.

Sentiment Analysis

Identifies emotions in text, photos, videos of faces (e.g., happiness, sadness, disgust, surprise).

Text Analysis

Provides ability to interpret and “understand” written or spoken text, and to respond with dynamically generated written or spoken text, in real time.

Natural Language Processing & Generation

Interprets speech input via phone or other channels, transcribing to text but also assessing topics and sentiment. Can recognize and interpret multiple languages.

Artificial Intelligence (AI)

The ability of a machine to perform cognitive functions we associate with human minds, such as perceiving, reasoning, learning, and problem solving. Examples of technologies that enable AI to solve business problems are robotics and autonomous vehicles, computer vision, language, virtual agents, and machine learning.

Machine Learning

Algorithms detect patterns and learn how to make predictions and recommendations by processing data and experiences, rather than by receiving explicit programming instruction. The algorithms also adapt in response to new data and experiences to improve efficacy over time. Allows the categorization of new input into known categories, as well discerning the unique groupings or categories hidden in large/complex data sets.

Personalization

The ability to personalize experiences and information to specific individuals with speed and efficiency so that the personalization can happen at scale. For example, learning scientists are beginning to develop software that keeps students in the “Goldilocks” zone so that the material is hard enough to be challenging, but not so hard that they are frustrated and give up.

Key Principles in Understanding How Humans Operate

Choice Overload

Have you ever looked at a supermarket display and felt paralyzed by the sheer number of options available? How about the menu at a diner? The Netflix homepage? Too many choices can be overwhelming, and lead to choice overload. When faced with a huge range of options, many people fail to choose the best option or fail to choose altogether. Having more options often leads to less realized choices.

By taking steps such as limiting the amount of unnecessary information presented, decreasing the number of choices presented, and increasing the meaningful differences between them, we can facilitate decision-making and prevent the paralysis created by choice overload.

Cognitive Depletion and Decision Fatigue

We often joke about being tired and hungry, or how we can’t function without our daily (or often twice daily) coffee. But fatigue is no laughing matter – being tired and hungry can deplete our cognitive resources and significantly affect our decision-making.

In order to combat the ubiquitous and pervasive effects of cognitive depletion, we must be aware of the extent of its effect on our decision-making and also design institutional structures so that decision-makers avoid making major choices when they are fatigued. Simple solutions like taking breaks and providing snacks to replenish decision-makers’ mental capacity can be effective ways to mitigate cognitive depletion.

Hassle Factors

Sometimes we don’t act in accordance with our intentions because of seemingly minor inconveniences, or “hassle factors”. We might intend to mail an important document for weeks, but the having to find a stamp and an envelope (or, heaven forbid, going out to buy them) can be enough to convince us that tomorrow is definitely a better day to mail that paperwork.

Minimizing hassle factors by eliminating unnecessary complications and confusing jargon, as well as providing clear channels for our actions are essential steps to help those actions match our intentions. When a process is by nature complicated or confusing, offering regular reminders and clear steps for seeking assistance can help people get through the hassle.

Identity

Many of our choices are impacted by our perception of ourselves and our roles. We all have multiple identities – for instance, someone can be a mother, a lawyer, a daughter, and a gardener – and each identity may carry different goals and values. Our perceptions, choices, and actions are often made in accordance with the identity (and its associated values) that is most salient to us in our moment of choice.

The existence of multiple social identities – and the sway they hold over our choices – means that the extent to which a message or option primes a specific identity can have important effects on the decisions that we make.

Limited Attention

Our ability to pay attention to several things at once is much more limited than we might think. This means that when our attentional capacity is stretched – for instance, when we are driving, listening to music, and texting all at once – we are left effectively blind to information we would easily notice under normal circumstances.

When we design products and programs, we assume that if people don’t choose them, then they must not want them. Often, though, people simply didn’t notice what we put in front of them.

Loss Aversion

Humans hate to lose more than they love to win. This common sentiment captures the main insight behind the psychology of “loss aversion.” More scientifically, we can illustrate loss aversion with a simple wager: Let’s say we flip a fair coin and tails means you lose $10. How much would you have to win from heads for you to be willing to take this bet? Economic theory would dictate that a tiny bit over $10 should be sufficient because your expected value from the bet would be slightly positive. However, most people demand $20 for heads in order to accept this gamble.

This aversion to losing is a powerful principle as framing something as a loss can sway our decision-making in many different domains.

Primacy Bias

Humans bias towards information that is presented first is called the primacy bias, and it can have serious and unintended effects.

Although the primacy bias is difficult to combat in all its forms, small steps like randomizing the order in which candidates’ names appear on the ballot can help us overcome it. We can also harness primacy to own advantage – for example, listing the item that is most critical for us to remember first on a to-do list. Doctors can make sure that they discuss the most important health information first.

Procrastination

We all procrastinate. We put off for tomorrow what we could (and often should) do today. These delays have real consequences.

The nature of the task can influence our tendency to procrastinate. We also tend to put off starting things that we cannot finish in one sitting. For instance, if a form takes 40 minutes to complete and we only have 20 minutes now, we will not bother starting. We’re more likely to procrastinate a complex or tedious task because it’s easier to be tempted by a more pleasant alternative in the moment. We can combat procrastination by making tasks easier, and breaking down complex tasks into bite-sized steps.

Social Norms

All of us are heavily influenced by our perception of what others are doing – the social norm. However, our perceptions are not always reality.

When most people are performing the desired behavior – but perhaps are not a silent majority – it is effective to make the descriptive norm more salient. Simply informing people of descriptive norm in these situation is often more effective than telling them what they should do (an injunctive norm).

Status Quo Bias

Beliefs, previous choices, and set routines form an individual’s status quo, which we tend to prefer to stick to. As a result, we often “choose” pre-set options even when many other options are available.

If we want people to be able to make unbiased choices, we must be careful not to set one option as the status quo. If there is already a status quo, we need to recognize that people will be less likely to switch to an alternative even if they would have chosen it in the absence of the status quo.

The Availability Heuristic

Estimating frequencies and probabilities is tricky. The easier it is for our mind to come up with examples of something, the more likely we are to overestimate the likelihood of the thing occurring. This is the availability heuristic – when we judge probabilities based on how easily examples come to mind.

There’s no “one weird trick” to get around our tendency to make use of the availability heuristic. When possible, try to make decisions based on what you know about “base rates” – that is, the average probability of a given event – rather than what seems likely given the examples you can remember.

The Planning Fallacy

We tend to be overly optimistic about our ability to finish tasks on time. We consistently believe the future will unfold as planned, and we rarely leave sufficient time to meet key deadlines.

To combat the planning fallacy, it helps to make salient the amount of effort involved in completing a task. This can be done by relating it to a similar past experiences or simply by informing people of the average time it takes to complete the task.