The Context of AI
The potential of artificial intelligence for growing businesses is irrevocable. With cloud, cyber-security, and big data, AI is considered as a top priority for technology investments in 2019. Converging more and more with other technologies in place, AI is driving the next generation of businesses by being at the forefront of almost every digital transformation. With 9 out of 10 executives from around the world describing AI as an important tool in solving their organizations’ strategic challenges, the rush toward the golden promise of exponential growth is real.
Despite being one of the most active advocates, the AI pioneer Yoshua Bengio himself recently issued a warning against the blind pursuit of this shiny new technology. Short-term thinking in AI development without any form of regulation and consideration of its potential systemic effects may lead to unforeseen implications in social, ethical and political arenas.
As the technology is still in its infancy, it’s hard to predict the impact in terms of job losses or displacement — McKinsey Global Institute suggests that by 2030, intelligent agents and robots could eliminate as much as 30 percent of the world’s human labor in a recent two-year study. Whatever the extent, it’s the responsibility of business leaders to ensure that its development is guided, at each step, by concern for its effects on humans. Before embracing it mindlessly as if it was a “silver bullet”, initiators of change need to make sure that they are solving the right problems where machines can bring real value, and doing so in a way that elevates human capabilities rather than displace them. Instead of looking at it as an isolated part of the organization which the core function is to optimize the process in place, we need to approach it as the new foundation of an interconnected set of elements that needs to be coherently balanced to truly bring meaningful value.
These long-term thinking and systemic mindset are not just nice to have in a time when agility and speed are the fundamentals of modern businesses, they are factors of success. Indeed, another recent study from McKinsey reported lack of strategy, before lack of capabilities, as the biggest challenge towards AI adoption.
“Iwould rather have questions that can’t be answered than answers which can’t be questioned” said the physicist and Nobel prize winner Richard Feynman. A strategy process always starts with questioning: questioning to help us gain perspective, questioning to overcome potential cognitive biases, questioning to make sure we don’t jump to the solution before we really understand the problem.
How might AI help you with the problem you are trying to solve?
What are the limits to what we can do with AI?
What tasks do humans do best vs machines do best?
In this time of rush and rapid change, never before has it been so important to take the time to pause. Considering what the effects of introducing AI to humans would do, is a good place to start to help define your overarching vision and the first steps to implementation.
Why this canvas?
Identifying how to use AI in order to provoke change in your organization or consumer experience is a complex challenge if you are starting in this field with no tools or established roadmap. The opportunities of applicability are wide and prioritizing where to take action first can feel overwhelming without a clear starting point.
Whether it’s your employees, your customers, or your external stakeholders, the intention of this tool is to invite you to reflect on your AI transformation from the point of view of the people who will be impacted first. Use it as a compass to guide you on this journey, by revealing the ripple effects of your transformation in the broader context of your organization and provoking conversations about the implications of designing a self-preserving, resilient and well-balanced AI system.
AI is bringing structural change to the way companies operate that expands far beyond technological considerations. Developing a design mindset with this canvas should help you anticipate how digitalization powered by AI will impact the constructing parts of your organization. Hopefully this canvas will also facilitate approaching AI as an interdisciplinary field, by creating a common language and serving as a starting point for discussion that expands beyond computer science and engineering disciplines.
With cultural and behavioral challenges being at the top of the list of the most significant barriers to digital effectiveness, inclusivity, transparency, and openness are indispensable requirements when designing new digital models. Companies have the responsibilities to educate and engage their employees to avoid any forms of ignorance that could raise uncertainty or misunderstanding about the value of AI, which I hope this tool will facilitate.
“To design is much more than simply to assemble, to order, or even to edit; it is to add value and meaning, to illuminate, to simplify, to clarify, to modify, to dignify, to dramatize, to persuade, and perhaps even to amuse.”
— Paul Rand
How to use it
This canvas is composed of 10 building blocks. Start by defining who will be the audience you are trying to solve a problem for (“Designed for”). It might be your customer or a specific department within your organization. Once you have picked the people whose perspective you will be doing this exercise for, you will be able to work your way through the canvas from the top-right to the left, ending with the three boxes at the bottom.
- Jobs-to-be Done
Similarly to Clayton Christensen’s Jobs-To-Be-Done theory, which inspired this tool, one of the goals is to ensure that you have deep understanding of the problem you are trying to solve, even before considering AI. In this very first section, your intention is to identify the problem that you want to address for your target audience.
- What are you trying to accomplish?
2. A.I. Promise
In this section, list the benefits of using AI in your organization for the task you are trying to achieve. Is it a process that will be streamlined? Will dormant data be cleaned and leveraged? Or will a specific step of your customer journey be enhanced? Consider the gains of introducing automation for the audience you are designing it for.
- How might AI help you with the problem you are trying to solve?
- Which benefits the use of AI might bring to your organization?
- How might the use of AI help you in achieving your goal or completing your task?
3. Benefits for Humans
Building upon what you wrote in the previous section, now list the pains that will be relieved for your human agents thanks to the introduction of AI. Whether its your employees or your customers, consider the pain points that will go away or be alleviated to allow them to be “more human”.
- How might the human agents part of your process benefit from using AI?
- What AI is freeing up in your human agents’ capabilities?
- Which repetitive tasks that your human agents are performing can be eliminated?
4. Machine Activities
In the same vein as the previous section, list the machine-only activities that will make the most of introducing algorithms to solve your problems. Are they activities which are too dangerous to be performed by humans? Is it spotting hidden patterns in complex datasets that are out of reach for human analytical capabilities? Or is it bringing speed and accuracy to perform a repetitive activity at scale?
- Which tasks in your process are best suited for machines rather than human?
- How might you make the best use of machine precision and accuracy?
- Which limitations of human abilities could be taken
over by machines in your process?
5. Human Activities
Now, list the exclusive human activities that your employees will keep doing. Consider the tasks that are best suited for humans rather than machines, requiring their judgment, creativity or experience. If humans totally disappear of the process because of the use of automation, consider activities like interrogating the output from the machine or overseeing the performance of the algorithm. There’s always room for humanity.
- What AI technology allows your human agents to focus on?
- How might you make best use of human agents’ judgment, flexibility and creativity?
- Which activities where AI can’t be used should be overtaken by your human agents?
“Your job in a world of intelligent machines is to keep making sure they do what you want, both at the input (setting the goals) and at the output (checking that you got what you asked for).”
― Pedro Domingos
In this section, we want to look at the hybrid activities where human and machines will work together to resolve your problem. Consider the opportunities where humans help machines and machines help humans. They work next to each other and interact, with ultimately the new technology increasing the productivity of your employees.
- What’s the role of your human agents in overseeing the performance of the machine and refining the model?
- How might you use the input from your human agents to improve the effectiveness of your machine?
- What is the relationship between your human agents and your machine?
7. Human Reinforcement
Ultimately, you don’t want your machine to perform tasks in disconnect with the rest of your organization. Its benefits should be transversal and allow you to amplify the capabilities of your human agents to unleash new levels of productivity at the same time. Look for opportunities to tap the full potential of a human-machine organization, where your people are able to boost their capabilities (or ‘human-ness’) by learning from the machine output.
- How might AI to influence the work of your human agents?
- How might AI enhance your human agents by adding to their capabilities
and improving their decision-making process?
- What feedback loop will you build between the machine and
your human agents to augment their intelligence and abilities?
8. Critical Thinking & Biases
Critical thinking is one of the many human cognitive abilities that will never be taken over by machines and that makes us truly unique. In this section, you have to consider the measures that you will implement to address the potential blind spots in your AI model. Identify where your people need to keep playing an influential role because it can’t be fully handed over to machines. You need to make sure you are truly attuned to how your AI works to avoid the ‘black box’ problem.
- What are the limits to what you can do with AI in your process?
- Which major or complex decisions can’t be taken over by machines in your process?
- How might you apply critical thought to your machine performance and overcome its potential biases?
9. Considerations & Implications
AI is a promising opportunity, but it also bears risks if it falls into the wrong hands. As our lives will become even more dependent on networked AI in complex digital systems, it’s important to pay close attention to the potential ethical, moral and social concerns that might arise. In this section, list the considerations that need to be tackled to ensure that you design a responsible AI that matches your values.
- How might you ensure that your AI is developed in a responsible, safe and reliable manner?
- How might you address the societal issues that the way you are using AI bring about?
- Which morals, ethical or philosophical norms should be taken into consideration?
10. Change management
As discussed previously, implementing AI in an organization may provoke profound changes and can’t happen without a clear action plan. To build trust and facilitate acceptance of the new technology, it’s critical that you take the time to consider how your processes will be reimagined, how you will communicate the value proposition of using machines-to your employees or customers-, and evolve the talents of your people. List the measures that you will put in place to successfully take the step toward the future of your organization powered by AI.
- How will you manage the anxieties of your employees that might be worried about their role being disrupted?
- How will you re-train your workforce to ensure a smooth transition of your company to the age of artificial intelligence?
- How will you communicate the difference between your human and machine-led experiences to your customers?
Here is a concrete example with a start-up I have been involved with, Mr Young. The company’s product is a conversational agent, powered by AI, to help its users cope with anxiety. Its objective is threefold: demystifying anxiety and mental health, providing a medium to evaluate and track anxiety, and helping get access to quick and personalized resources to get in control. Through conversational platforms, Mr Young provides screening tests, suggests videos and articles to CBT-based exercises, or connects directly with a mental health professional.
The tool is aimed at human resources departments, with the promise of allowing members of their organization a quick access to multiple valid solutions in order to look after themselves.
In this case, the audience we are designing the canvas for is the HR department, who is considering implementing an AI-assistant to create a work environment with stable well-being.
Transforming organizations by using new technologies is a challenging task, especially when clear blueprints do not yet exist. This canvas won’t provide you with all the answers, but I hope it can play the role of a starting point in your journey. As humans, it’s our duty to design systems that celebrate and elevate our human qualities.
Creating AI legitimacy requires us to slow down and engage in an open discussion about its implications. The stakes are too high for embracing the ‘move fast, break things’ rhetoric. Ultimately, including your people and getting their feedbacks on your vision will help to change the negative perceptions of AI fueled by pop culture, but also help you to explain the potential outcomes of embracing the new technology. It’s time to change the topic from which jobs will be eliminated to how AI should be implemented to enhance human capabilities. To stop thinking about robots as human replacement but start thinking about them as partners with a different skill set.
In these endeavors, human-centered design — and its promise to build new solutions that are tailor-made to serve the needs of the people it’s being designing for — is an essential tool for change makers who aspire to create healthy foundations for a viable future where humans are truly appreciated, and not slaves to the machines.
“We think too much and feel too little. More than machinery we need humanity. More than cleverness we need kindness and gentleness. Without these qualities, life will be violent and all will be lost.”
* * * * *
I’m very interested to hear about your thoughts and what happens when you use this new canvas for Human-Centered AI. I hope you will experiment with it and tell me how it’s working for you, and how it might be improved. Please share your thoughts or feedback by leaving a comment.
Click below to download a PDF version of the canvas.