Here at KUNGFU.AI, we have a passion for not just finding and building fantastic AI strategies and solutions for our clients, but also ensuring that AI is a positive force in the world. In practice, this explicitly means that we have weekly robust internal conversations about ethics, our values, and governance. Furthermore, we maintain standards about the type of work we are willing to do, from an ethical standpoint, and make a point of aligning the ethical message between our sales, marketing, and engineering teams. With this in mind, we have done a lot of internal talking and would like to take what we have held until now as a bit of a “private value” and make it public and explicit.
Therefore we would like to announce that KUNGFU.AI has recently signed onto the Safe Face pledge. Concordant with the spirit of this pledge, we are now outlining how we feel about it, and how we wish to implement it. The explicit planks of the pledge are listed below, pulled directly from their site:
With this in mind, here is how KUNGFU.AI plans on implementing these points.
First, everything starts with the fact that KUNGFU.AI maintains an internal ethics discussion group that meets weekly, including leadership, actively seeks outside voices, and coordinates with sales and marketing regarding ethics concerns regarding current projects and potential future projects and contacts. We actively engage in robust conversation regarding ethical issues and have walked away from contracts that do not match our principles.
This commitment extends to us choosing clients and partners that align with us not just in a business sense, but in an ethical sense. While the nature of our business as a professional services firm does at times prevent KUNGFU.AI from directly publishing the details of models delivered, what we will do is work with our clients to determine that they have the correct processes in place to evaluate and correctly manage any models that they deliver, and that they stay compliant with our anti-facial recognition pledge and have processes in place to address any undue discriminatory bias. While some may see this as a potential source of conflict, we, in fact, see this as a core part of our added business value, which involves not just delivering fantastic AI solutions, but also giving clients the ability to understand what we have done, expand on it, and integrate it into their systems.
Furthermore, KUNGFU.AI hereby commits to not engage in any AI-based biometric (this means faces, but it also means voices, fingerprints, etc.) recognition of individuals, in addition to our longstanding commitment to not attach any of our products, research, or solution to devices intended to injure humans in any way. In our internal discussions, it has also become clear that we need to just internally discuss this, but also include a prohibition on repurposing our software for these purposes into our future contracts. While we are committing to this, we are also leaving one explicit caveat to this pledge -- are explicitly leaving open one form of facial and biometric recognition -- developing a model that answers the lowest level question of facial recognition -- identifying whether a collection of pixels contains a face at all. The reason for this is that in order to guarantee that a model does not inadvertently use facial data as a feature, it may be necessary to develop technology to redact faces from input data. By broadly refusing participation in mass-surveillance schemes, those of us at KUNGFU.AI hope to both realign norms in the industry against participation in them.
Finally, our intent is to make our participation in all of this public. By discussing these matters in a public forum, we hope to both help change the norms in the technology industry as well as provide ourselves as a resource for addressing these issues to both technical and non-technical audiences. So please, watch this space, or contact info@kungfu.ai if you have any questions regarding the Safe Face pledge, our compliance to it, this document, facial recognition, or our ethics practices in general.
