Our Ethical Principles
Receptiviti’s mission is to help companies create and maintain people-first workplaces where culture is prioritized and recognized as the key driver of organizational health and overall business performance. We believe that doing so requires a fundamental shift in the way that culture is understood and measured.
Most modern organizations have well-established platforms for measuring key performance indicators like sales, churn, and margins. Culture and organizational health, however, have traditionally been hard to measure, and the data and tools used to measure them have been crude at best.
Data integrity has long been a key challenge to accurately measuring culture and organizational health. Workplace surveys, for example, are often undermined by bias and limited by what employees are willing to share, leading most organizations to make critical decisions based on flawed data. Solving this problem requires new methods for data collection and interpretation that yield valid and explainable insights about the human experience in the workplace and organizational health.
Valid insights begin with data integrity. First and foremost, we believe that collected data must possess the following five characteristics:
Accuracy and Validity: Data is reflective of reality, both quantitatively and qualitatively.
Reliability and Consistency: Data is gathered in such a way that each data point reflects the same information defined in the same way.
Contextual Relevance: Data is socioculturally, linguistically, temporally, and regionally relevant to the insights derived from it.
Completeness: Data contains all information vital to informing the insights derived from it, and data contains all entries needed for insights to be unbiased and accurate.
Appropriateness: Data directly reveals the information being derived from it without being stretched or misconstrued.
As decision-making systems become increasingly sophisticated, concerns about ethical implications have increased in tandem. We recognize that, whether by design or as a result of the complexity of deep learning systems, artificial intelligence and algorithmic decision making can lead to bad decisions that can have severe and detrimental impacts on people’s lives.
To ensure that Receptiviti is utilized ethically and as intended, we have established ethical guidelines that direct our business practices, govern our research, and serve as a foundation for the products we build. We commit to upholding these guidelines in every facet of our work, and we will consistently revisit them as we continue to grow and learn:
1. Protect privacy
Due to concerns about invasiveness, research indicates that 30 percent of companies currently consider it a “risk” to introduce workforce analytics software. Unfortunately, companies that avoid workforce analytics forego the opportunity to gain valuable insights that actually reduce risk within their organization.
To ensure that our customers and their workforces derive the full benefit from Receptiviti, we have engineered privacy and anonymity as core structural elements of the platform. Our hybrid cloud architecture ensures that a customer’s data never leaves that customer’s technology environment. Names, identifiable information about people, and any content generated by people are never retained by the platform, nor can this data be traced back to individuals. Further, we are continuously working to comply with the ever-changing regulatory and compliance landscape, whether it be universal standards like GDPR or an individual customer’s unique information security requirements.
2. Uphold scientific rigour
We pride ourselves on the scientific depth of our platform, and we are constantly working to expand and improve both our science and technology. A core element of the Receptiviti platform and science is LIWC, an approach to natural language processing that exists at the intersection of language and social psychology, and has been continuously developed over the past 30 years by Receptiviti co-founder Dr. James W. Pennebaker. Receptiviti maintains the exclusive commercial rights to LIWC, and we take our role as its custodians extremely seriously.
Our team is comprised of experts, PhDs, and practitioners with extensive training and experience in linguistics, social psychology, machine learning, and statistical research methodologies. We use meta-analyses, proprietary experimental research, and our wide and ever-expanding range of datasets to construct and validate our capabilities, and we continue to build our network in academia to remain abreast of new developments in our field and adjacent areas of research.
3. Avoid reinforcing biases
A pressing concern related to both AI and workforce analytics systems is the potential for such systems to reinforce workplace biases. We are cognizant that our technology can either decrease or exacerbate bias depending on how findings are interpreted. To ensure that insights are correctly understood and utilized, we seek to educate our users directly through the platform. The prevention of bias reinforcement is, and will continue to be, a key and critical priority for Receptiviti.
4. Maintain explainability
Public policy experts across the globe are petitioning governments to regulate the use of black box algorithms based on the dangers of allocating decision-making power to opaque technologies.
Our algorithms are not black box algorithms, meaning that we can “unpack” them to understand how insights are generated. By maintaining transparency, we provide our customers with expandable insights to support their decisions and we insulate ourselves from the impact of future legislation in this area.
We believe that a sustainable approach is one where such systems generate explainable insights that serve only to inform human decision-making, and where responsibility for the decisions and their implications reside solely with the human decision-maker.
5. Maintain transparency about what Receptiviti can and can’t do
We believe in maintaining transparency with respect to the limitations of both our science and our platform. We will not present or portray Receptiviti as being capable of doing things it cannot, and we will maintain this level of transparency as we continue to learn and expand the platform.
6. Follow a “first, do no harm” principle
In everything we do, we seek to minimize any possible risk that our technology can be used for harmful purposes.
7. Incorporate feedback
We actively solicit feedback, both directly through customer conversations and programmatically through the platform, to identify any possible inaccuracies and to understand how insights are being utilized. The information we gather is used to educate ourselves, expand our science, improve the capability and validity of the platform, and to support our ongoing efforts to continuously reduce the potential for harm.
Applications we will not pursue:
In addition to the guiding principles outlined above, we have identified specific applications of our technology that we will not develop, endorse, or support:
Applications that are likely to cause harm: While defining applications as harmful can be a subjective exercise, we believe that if there is a material risk of harm, we will proceed only where we believe that the benefits substantially outweigh any risks, and we will incorporate any appropriate safeguards necessary to ensure the platform continues to operate within our ethical boundaries.
Applications that violate internationally-accepted norms, or whose purpose contravenes widely accepted principles of international law and human rights: We will not support any use of the platform that involves analysis of non-public language data or whose subjects are not made aware that the technology is being used, and we will not work with any party or customer who conducts analysis of this sort. Further, Receptiviti will not work with any entities on initiatives relating to government surveillance of citizens or other unknowing parties, groups, or populations.
We recognize the power of our technology and the importance of input and feedback from our customers, our advisors, and the broader community that uses or is impacted by our platform. As Receptiviti’s capabilities continue to expand, we will continue to work with stakeholders to promote learning, public discourse, and thoughtful leadership in our domain. We will endeavour to share our findings and incorporate our learnings as we work to continuously improve our science and our capabilities.
We welcome your feedback.