As technology continues to weave its way into our lives and natural environment, there's a growing need to ensure that it’s developed and used responsibly. This blog shares some lessons for creating ethics and responsibility frameworks, including defining what's meant by an "ethics framework” and setting realistic timelines.
The importance of ethics
I recently attended a talk where a senior machine learning engineer from an influential technology company admitted that up until Chat GPT-4 launched, he used to roll his eyes at the very mention of tech ethics. Speaking with him at the event, he said the thing that changed his mind was how powerful AI tools had become. Suddenly, he could see the need for oversight and guardrails because the ‘what if’ scenarios had grown in number and scope, seemingly overnight. There are many who feel the same way.
In March 2023, the Future of Life Institute published a letter calling on AI labs to immediately pause training AI systems that were more powerful than GPT-4. To date, there have been 33,709 signatories. While the letter didn’t result in the recommended pause, the media coverage the letter received played a role in causing the first domino to topple over. Tech CEOs were summoned for questioning, the UK held the first AI Safety Summit which world leaders and tech elite cleared their schedules for and the US – a country not known for regulating big tech – published a “blueprint for an AI Bill of Rights”. The other change that has happened is that more organisations seem to be funding, or are looking to fund, ethical and responsible AI workstreams.
Conversations about frontier technologies, like AI, most often focus on the components needed to build and deploy. This tech-centric lens often masks the complexities needed to do so responsibly. Frontier technologies bring with them frontier ways of working and frontier skill sets. Ethical thinking and considerations or how to design technology responsibly and with moral imagination - whatever you want to call it - is a frontier skill set.
The benefits of creating ethical & responsible technology frameworks
Minimising harm, controlling risk and maximising positive impact
As much as technology has changed our lives for the better in some ways, we have to be honest about the negative consequences that have occurred too; from deep fakes that can be used in disinformation campaigns and in-home IoT devices that can be controlled by coercive partners, to personal account takeovers. Ethical and responsible technology frameworks provide tech teams with tools, guidelines, methods, and new ways of working that supports collaborative deliberation, critical thinking, and new design patterns that can be implemented to minimise potential harms, reduce risk, and amplify positive outcomes.
Building trust through accountability and transparency
The Centre for Data Ethics and Innovation Public Attitudes Tracker Survey monitors how public attitudes towards the use of data and AI vary over time. In the latest findings, published on 2 November (2023), it was found that trust in big technology companies to act in people’s best interest had fallen since 2021. Particularly concerning is that this belief is most likely to be held by those who are most positive and knowledgeable about technology. When made transparent, baked into an overarching governance structure and managed by an accountable individual/s, ethics frameworks can help to build trust with colleagues, customers and partners. It signals that you understand the importance of creating and implementing tech in a responsible way and that you take it seriously. It can become a strong marker of responsible corporate citizenship.
Becoming a more resilient organisation
Over time, tech ethics and responsibility frameworks can help an organisation become more resilient to a changing regulatory and legislative landscape. Very often, tech ethics is the experimental grey space where we test the boundaries of right and wrong, and what we should or shouldn’t do in a controlled and low risk setting. It’s very often through this exploration of testing, learning and sharing that new ideas, arguments and case studies emerge. These ultimately become the evidence base that lawmakers use to build legislative frameworks. Organisations who build and implement ethics framework are better positioned to comply with future legislation because chances are, they are already doing the right thing.
5 lessons to help you create tech ethics frameworks
1. Define what you mean when you say “ethics framework”
This might sound blindingly obvious, but the obvious is often overlooked. There is a really good chance that when the phrase “ethics framework” is said, that everyone is thinking of a slightly different thing. For some, a framework could be a set of values and principles and an accompanying guide for how to implement them. For others, it could be a design process and compendium of methods and tools that guide AI teams to imagine, design and deploy in a way that is defined by the organisation as ‘good’. For a different group, an ethics framework needs to include all these things. Before starting on the work, define, scope and collectively agree on what it is.
2. There is no ethics drive-thru
Typically, ethics is contentious. This is because we all define ‘good’ in a slightly different way. Realise this going in, and acknowledge this is not something that can be spun up overnight, or dare I admit, in a few sprints. It requires conversations, many of them. You’re going to go through more than one iteration. It’s going to take time and energy. Why? Because an ethics framework has to represent an organisation - one that is likely quite complex, and has existing processes in place and is made up of many people. If you’re taking the lead on creating an ethics framework, remember this isn’t your ethics. It belongs to, and represents, a collective. And building collective understanding and consensus takes time.
3. Determine who needs to give feedback, and who needs to sign off outputs
There is the temptation with this kind of work to open it up too widely for input. Ethics and responsibility frameworks should be treated in a similar way to other projects. Create a stakeholder map and plot who needs to be kept informed, who needs to contribute their feedback and who needs to approve the work. While the ethics and responsibility framework will belong to an organisation, trying to include everyone’s feedback is an impossible task that will only result in unnecessary delays. Creating an Ethics Board that is made up of a diverse range of colleagues from across the business can be a great way to consult and listen to a multiplicity of voices outside of the C-suite.
4. Identify where ethics will ‘live’ once it’s been created
For many – if not most – organisations, ethics and responsibility frameworks are very new and so that begs the question, where on earth does this now live in the organisation? Once you’ve created it, who owns it? Who will update it? Who will ensure that teams are indeed following guidelines? At first, you may think that your compliance/assurance team will be a natural fit, but are they well staffed enough to take this on? Also, do they have the knowledge of how to assess whether ethical guidelines have indeed been followed?
Alternatively, the Product or Delivery Leads could be responsible for prioritising these new tasks and they could work closely with compliance/assurance teams to show how ethical considerations have informed design and implementation decisions. In this way, ethical considerations become a control function to manage risks. There is no one-size-fit-all solution here, but rather is dependent on the set-up and capacity of teams and departments.
5. Develop associated training
As mentioned above, frontier technologies bring with them frontier skill sets. Like all new skills, people don’t develop them by osmosis. They need the right environment, training style and time to learn and practice. Great training programmes usually give everyone the opportunity to learn the basics of an essential skill but then allow employees to branch off into discipline-specific training that makes most sense to them. In this way, everyone learns the 101 but engineers get to focus specifically on how to apply different technical treatments of fairness and user researchers are trained in how to interrogate data sets in order to identify which users are represented and which aren’t.
Conclusion: ethical technology matters
Work relating to the ethics and responsibility of technology is both hugely rewarding and beneficial. People often asked me why I care so much about this work and my answer is always the same. Technology, and by extension data, are increasingly determining what it means to live a good life. Technology isn’t just about apps and services and clicky buttons. It’s bigger than that. It’s about ensuring that we help people live good lives and as an industry, that’s worth money and time and good old fashioned effort.
We're experts in building capabilities as well as remarkable AI solutions that help people lead good lives. Learn more about how we help organisations like yours to reshape operations, products, and services like never before through the use of AI.