From the AI Global Council: 9 AI questions leaders should be asking to avoid pitfalls

Artificial intelligence (AI) has continued to emerge as a transformative force in our rapidly evolving technological landscape. To participate in the advancement of AI technologies, policies, and ethics on a global scale, the AI Global Council—a group founded with the vision of fostering collaboration and reshaping the responsible development and deployment of AI—is addressing important questions regarding AI adoption. The first topic on the council’s mind: What questions should leaders be asking about AI?

According to council members from educational institutions and organizations such as Credera, Mastercard, AWS, McDonalds, United States Patent and Trademark Office, GoGuardian, and Pluralytics, there are many questions leaders should be asking about AI.

Watch the council's full conversation and read key takeaways below:

Nine questions to ask before starting an AI project

1. What’s the outcome we’re trying to achieve?

Articulating the desired outcome is key prior to landing on an AI-based solution. After spending enough time considering their desired outcome, leaders should ask “How are we going to measure that this outcome has actually been achieved?” Defining how to measure success ensures the chosen solution will be effective. The next step after success definition is to implement the AI solution in small chunks to get rapid results.

2. Do you have the necessary resources, including the required data, talent, and culture?

Organizations must have the essential resources to implement AI solutions or tools in a quick, agile manner. Key resources include data, talent diverse enough to bring different perspectives, and a culture that fosters fast-paced, agile ideation and delivery.

Depending on how quickly leaders are implementing AI, an acquisition may be the best solution to address a lack of necessary resources. Executive leaders should be asking their teams three questions about resources:

1. Do we have it?

2. Do we need to source it?

3. How can I help you go fast?

It should be noted, these three questions are important to ask all types of teams, not just technology teams.

3. Do we have a process for this collaboration?

Having a pre-existing process for innovation in place makes implementing AI much easier for leaders and their teams. Without having a fast-paced, iterative innovation process for collaboration, the right strategy and technological and business resources may not be enough for leaders to implement an AI solution or tool.

The cross collaboration that implementing AI requires often involves disciplines including data, technology, and business strategy. To correctly weigh risk prior to and during the implementation process, collaboration must occur between all these disciplines and specifically between individuals with experience in data risk such as privacy professionals and legal professionals. These key roles must be at the table when AI implementation is being discussed so they can be part of the iterative process to move fast.

4. Does our culture support transformation?

Especially when considering new technologies such as AI, having a thriving, motivating, and propelling culture that supports pivots and transformations is key. To implement AI successfully, leaders must remember that culture can stifle or accelerate their efforts. Handling failures and moving forward successfully is incredibly important in forward-leaning spaces like AI, and everyone should get to take part in the pivot to a new AI tool.

5. Where is experimentation already happening within our organization?

To understand where AI can help their organization, leaders should look to understand where employees are already experimenting with AI. Experimentation with AI is oftentimes happening in unexpected places, and leaders should look to find out which problems people in their organization are already trying to solve with AI.

These seeds of experimentation could very well grow into the biggest opportunities. Because experimenting is already happening, leaders should also consider putting guardrails on confidential data and defining what is in bounds and what is out of bounds regarding AI.

6. Is AI the actual solution?

When leaders consider the problems they’re trying to solve, they need to consider if AI is the actual solution or if there are other processes that should be leveraged. The first step in answering this question is looking at what processes, technologies, or capabilities the organization is already building upon that could be used to address the problem at hand. Another key consideration for leaders to make is thinking about what the institutions, attitudes, beliefs, and norms in their organization are surrounding value generation. For example, understanding that your organization has a risk-averse culture provides substantial direction in deciding whether to go with an AI solution or not.

7. The process gives us an output, but is it the right outcome?

For generative AI and AI in general, the output is different from the outcome. To understand if the AI solution or tool is providing the right outcome, leaders need to implement a feedback mechanism that captures the user’s perspective on if the service is working or not. Closing the feedback loop is key, and data on the process’s outcome for the user will help leaders understand if the AI solution is addressing the right problems.

8. How sensitive is the result we are trying to achieve?

More sensitivity in the outcome means more need to examine the difference between output and outcome. Instances where more distinction between output and outcome is needed include scenarios when the output is going to be deeply personal to an individual. For example, if the output will make a card decline, affect a screening for a secure network, or hurt a customer’s credit score, then the level of sensitivity in the desired output would be high.

9. Where can I use this technology to find patterns I didn’t expect?

AI technologies can be used to uncover unknowns and solve specific problems. To make full use of AI in their organization, leaders should consider how they can apply this technology to find new insights and address curiosity while considering risks.

Moving forward with your AI initiatives

Asking the right questions is key when considering new AI technologies. When considering machine learning solutions, leaders must consider how AI fits into systems or products in their organizations rather than thinking of AI as a silver bullet to fix all their problems or automate processes.

Implementing AI requires a human centered approach that considers the input from machine learning algorithms to create an output that delights customers and aligns back to the problems leaders are aiming to solve. To develop and deploy AI-based solutions, leadership must deeply consider where and how AI fits into their organization rather than fixating on it as a shiny end goal.

Explore more about the AI Global Council

The AI Global Council is focused on taking amorphous and challenging topics and sharing a balanced and holistic perspective that can move everyone forward.

To explore more about the AI Global Council, follow along on LinkedIn or reach out at findoutmore@credera.com.

• Data Science, Artificial Intelligence & Machine Learning
• Artificial Intelligence
• AI Global Council