A Practical Guide for Building Ethical Tech

“Techlash,” the rising public animosity toward big tech companies and their impacts on society, will continue to define the state of the tech world in 2020. Government leaders, historically the stewards of protecting society from the impacts of new innovations, are becoming exasperated at the inability of traditional policymaking to keep up with the unprecedented speed and scale of technological change. In that governance vacuum, corporate leaders are recognizing a growing crisis of trust with the public. Rising consumer demands and employee activism require more aggressive self-regulation.

In response, some companies are creating new offices or executive positions, such as a chief ethics officer, focused on ensuring that ethical considerations are integrated across product development and deployment. Over the past year, the World Economic Forum has convened these new “ethics executives” from over 40 technology companies from across the world to discuss shared challenges of implementing such a far-reaching and nebulous mandate. These executives are working through some of the most contentious issues in the public eye, and ways to drive cultural change within organizations that pride themselves on their willingness to “move fast and break things.”

WIRED OPINION

ABOUT

Zvika Krieger is the World Economic Forum’s head of technology policy.

From these learnings, we’ve distilled practical advice for corporate leaders to develop an effective approach to mitigating or preventing negative impacts from their products and regain public trust.

Start in the trenches

While accountability for harmful products often happens at the executive level, decisions that lead to them are often made by engineers and developers on product teams. If you look at the recent tech scandals, from discriminatory advertising to proliferating hate speech, most of them did not involve a pivotal moment when someone decided to proceed with a product despite knowing how it could be abused or misused. Rather, they usually emanate from an unconscious design decision that had unintended impacts.

It’s natural that most tech developers have a bias toward imagining the ways their products can benefit society. To counteract this, employees need tools to think beyond the most evident use-cases; help predict a range of harms, from bias and discrimination to tech addiction to emboldening extremists; and develop strategies to mitigate those outcomes. EthicalOS and DotEveryone offer toolkits that a number of ethics executives have successfully used to this end.

Don’t reinvent the wheel

Identifying red flags is just the first step. There needs to be a process to ensure that those red flags are raised to an appropriate level of seniority and adjudicated transparently and consistently. Some ethics executives experimented with creating a new process for these “ethics checkpoints,” but quickly realized that this unduly burdened the notoriously tight product development cycle, or was ignored altogether.

What’s proven to be more effective is piggy-backing on processes that are already well-engrained in the product development roadmap, such as those that have been created in recent years related to cybersecurity, environmental sustainability, and accessibility. This allows straightforward concerns to be addressed quickly, and more complex or sensitive ones can be escalated for deeper review. The more you can make an engineer’s life easier, the more likely your ethics review process will succeed.

Design for scale

As tempting as it may be to see a new “ethics office” as the panacea for a company’s problems, ethics executives quickly realized that they were unable to keep up with the demands for support from across the company, no matter how big their new department grew. Devoting your whole ethics teams to dive deep on a few controversial topics or complex new products for a few months is useful to initially hone a methodology. But that approach doesn’t scale when there is a need for attention and consideration across all products and features.

Instead, companies like Microsoft are now finding success with training “ambassadors” or “champions” embedded in teams—usually a role they play in addition to their regular jobs—to bring their teams heightened sensitivity toward unintended impacts, as well as to help their teams navigate raising flags and escalating concerns. Empowering people within teams ensures that they have the contextual intelligence as well as pre-existing credibility needed to be trusted and effective.

It’s all about the culture

You can create the most well-designed process, but no one will follow it (or they will turn it into a superficial box-checking exercise) if they aren’t incentivized to do so. The priority for most engineers is to ship their products fast. That is how they are primarily evaluated, and that is what the culture prizes in most tech companies. To get real about responsible innovation, companies need to build these practices into individual and team KPIs, objectives, and performance reviews, as well as criteria for promotions, raises, bonuses, and even hiring.

These hard incentives need to be complemented with a range of soft incentives. Think about how your company celebrates a new product or feature launch. Maybe the team is congratulated over email or at the weekly all-hands meeting. How can you do something similar for when a new product isn’t launched (or has been significantly tweaked) because an ethical concern was surfaced? Employees have a keen sense of what is valued in an organization, and ethics executives are seeing that subtle cues like these can go a long way in changing behavior.

As the world’s leaders convene in Davos, Switzerland, for the World Economic Forum’s Annual Meeting, numerous sessions will address the challenges and opportunities presented by “techlash.” Executives may be hesitant to discuss their nascent ethics initiatives. But the tech industry does not have the luxury of experimenting within closed company silos until these processes, toolkits, and training curricula are perfect. Ethical lapses in one company impact public trust in technologies themselves. The only way to operationalize a vague concept like “ethics” is to share what’s working and what’s not—in real time—so the industry as a whole can begin to rebuild trust.

WIRED Opinion publishes articles by outside contributors representing a wide range of viewpoints. Read more opinions here. Submit an op-ed at opinion@wired.com.

More Great WIRED Stories

Read More