AI technologies are making continuous advances in domains like industrial robotics, logistics, speech recognition and translation, banking, medicine and advanced scientific research. But in almost every case, the cutting edge AI that drives the advances drops from attention, becoming almost invisible when it becomes part of the overall system.
The fact that most AI use today is invisible can lead to the erroneous assumption that it is not delivering on expected value, and this can translate to caution when considering new ways of applying AI to business operations. We have also seen some in the executive business ranks to treat AI as a buzzword not worthy of focus. This is probably another outcome of AI becoming invisible the more it is mainstreamed.
If it is your competition becoming apathetic about AI they may be doing you a favor. To keep your organization for doing them a favor back, we provide our recommendations for your approach to an AI strategy here.
Our most strategic recommendation regarding AI and your business is to ensure the executive leadership team is engaged and owns the strategy. Setting a strategy and leading to optimal outcomes is the task of your senior leadership team. Staying informed on the state of AI and lessons learned from other businesses will help you do just that.
This is the most important of our four recommendations for business leadership of AI. Our full list is:
- Ensure leadership engagement and set the strategic vision
- Organize for success
- Think comprehensively about data, algorithms, infrastructure, technical talent and risk
- Build in AI security, and seek independent evaluation of your AI risk posture
Ensure leadership engagement and set the strategic vision
Lead with informed AI vision for the enterprise
- Establish vision and create measurable goals
- Crawl, Walk, Run
- Use R&D to prototype and iterate
- Use open development and leverage community lessons learned
- Ensure oversight
- Build in independent evaluation
- Establish a culture of DevSecOps to address security concerns during the design and implementation process.
A mistake we have seen too often is for major AI activities to be relegated to the IT team. IT is critically important, of course, but it is the corporate leadership team who should set the vision and ensure the strategy is in place to accomplish business objectives. The entire leadership team should be involved.
A key requirement the leadership team should set is the organizational appetite for AI transparency and the explainability of results. Depending on your business, you may want all algorithms to be totally transparent and all results to be explainable. More on this aspect of risk is articulated below.
It is leadership who will also determine how comprehensive the AI projects should be. As you lead your team in articulating the vision, be sure to consider the many successful use cases already established in the community, as well as the many problems already identified and addressed. If you are building a “moonshot” program, you may well be setting yourself up for failure.
Organize for success
Consider the structure of your AI teams and their reporting structure. You basically have two choices: Centralize or Decentralize. The right decision will be one that fits the business objectives of the firm. In deciding reporting structure, consider expected outcomes.
The decentralized approach allows your line of business managers to have maximum control their own AI projects. This can result in projects with strong support inside the various units of your business, but is generally going to be slower to deliver and more expensive. Centralized approaches allow for creating of in-depth knowledge and a greater depth of expertise. The reporting structure of a centralized AI organization may be to the COO or CEO.
Any team will require technical experts, of course. But remember this is as much art as science. Your approach to AI will need business leaders, and probably a good mix of staff from functions like legal, HR, marketing, and IT.
Think comprehensively about data, algorithms, infrastructure, technical talent and risk
Every AI solution requires four things:
- Technical Talent
Additionally every AI solution comes with significant risk to the firm.
It is the job of corporate leadership to ensure AI solutions are appropriately optimized in all four of those areas.
Build in AI security, and seek independent evaluation of your AI risk posture
There are things executives can, and should, do to reduce the risk that comes with AI solutions. The most important step is to realize the role of the executive. Leaders are expected to ask the right questions and make sure policies and processes are in place to achieve results. This is especially true in AI security.
We outlined many challenges in AI security above. In meeting these challenges, ensure your team is thinking about risk reduction comprehensively. This will include:
- Establishing a secure foundation: AI security is much more than just IT security for AI platforms, but starting with a solid IT security program is a critical first step. This includes using protected infrastructure that is well managed and monitored, protecting data from unauthorized access and manipulation, and ensuring robust access controls for all components of the AI solution, including algorithms. But this is only the beginning.
- Ensure unbiased, protected data: No matter what the type of AI is deployed, data is key. The better you understand the data the better you will be able to ensure your results are optimized to deliver value. This includes helping to mitigate bias and helping to detect adversary action. Problems to check for include incomplete data, errors due to super distant outliers, missing data, duplicate data or areas were collection was from an untrusted source. You will want to protect the data from unauthorized access and unauthorized changes, but will also need to test it for bias.
- Assess training data: Conducting assessments on training data used to initialize AI solutions to reduce the likelihood of manipulation.
- Build in algorithmic transparency: By ensuring your AI solutions team is aware of your requirement for AI transparency, solutions can be built that let you know how results are being achieved. This will help make the AI solutions and their results more explainable. Other AI systems are so complex that being transparent is not going to help. Deep learning systems change via back propagation, for example, and over time can become so complex no human can understand them. They become inscrutable. In these cases, understand the strengths and weaknesses of the algorithms.
- Monitor: Monitoring during execution will help your team ensure checks are run for bias and other problems in results.
- Establish AI incident response capability: Design to minimize risk but be ready for surprise. Incident response may include reverting changes to algorithms or training data. However, incidents of significance may involve coordination and communication to internal and external stakeholders.
- Track the regulatory environment: Companies also need to closely follow changes to the regulatory environment. The GDPR, for example, requires the right to an explanation for individuals affected by computer-based decisions.
- Ensure independent assessment: Businesses should always seek to have their AI solutions independently assessed for the likelihood of bias in results. AI results should be fair and treat all people equitably.
No matter what your line of business you can use enhanced automation and better artificial intelligence to serve your customers and improve your organization’s ability to compete. Doing so requires leadership and an understanding of both the capability and risks of AI. The insights we provide in this reference are meant to give you a leg up in your AI projects, but this is only a beginning. New lessons are continuously being learned in AI, which means we are all being treated to a great opportunity for continuous learning in this field.
Change in the technology landscape, in geopolitics, in business innovation and consumer buying patterns will continue and paces impossible to measure. Fortunately, you don’t have to measure the pace of change. What is required is an ability to spot changes relevant to your business and markets and take appropriate action.
There is a model for doing this. The OODA Loop. The Observe – Orient – Decide – Act process is a framework that can help you spot opportunity and risk in an age of continuously accelerating change. Applying OODA to your business can help you assess what change means to your employees, suppliers, customers and market.
The essence of what you need to know about AI is that success requires leadership, and leaders should maintain a fluency in the key concepts articulated here. We’ll continue to track the megatrend of AI over the course of the year and will be hosting special conference calls and private events for our OODA Network members.
More on Artificial Intelligence:
Michael Kanaan has helped a wide swath of decision-makers better grasp the nature of AI. He has a knack for expressing complex topics in clear, accurate and succinct ways and many of us in the national security community have already had the pleasure of hearing from him in person or in conferences. His book, T-Minus AI: Humanity’s Countdown to Artificial Intelligence and the New Pursuit of Global Power, provides context and insights in a way that can help concerned citizens and business leaders better grasp the issues of AI. He gives us all a call to action to learn more because as he makes clear in the book, the countdown to AI is actually over.
This is the only security framework we have seen that helps prevent AI issues before they develop
This plain english overview will give you the insights you need to drive corporate decisions
By studying issues we can help mitigate them. And there is, unfortunately, a long history of AI going wrong to help us map out mitigation methods.
Black Swans and Gray Rhinos
Now more than ever, organizations need to apply rigorous thought to business risks and opportunities. In doing so it is useful to understand the concepts embodied in the terms Black Swan and Gray Rhino. See: Potential Future Opportunities, Risks and Mitigation Strategies in the Age of Continuous Crisis
Explore OODA Research and Analysis
Use OODA Loop to improve your decision making in any competitive endeavor. Explore OODA Loop
The greatest determinant of your success will be the quality of your decisions. We examine frameworks for understanding and reducing risk while enabling opportunities. Topics include Black Swans, Gray Rhinos, Foresight, Strategy, Stratigames, Business Intelligence and Intelligent Enterprises. Leadership in the modern age is also a key topic in this domain. Explore Decision Intelligence
We track the rapidly changing world of technology with a focus on what leaders need to know to improve decision-making. The future of tech is being created now and we provide insights that enable optimized action based on the future of tech. We provide deep insights into Artificial Intelligence, Machine Learning, Cloud Computing, Quantum Computing, Security Technology, Space Technology. Explore Disruptive/Exponential Tech
Security and Resiliency
Security and resiliency topics include geopolitical and cyber risk, cyber conflict, cyber diplomacy, cybersecurity, nation state conflict, non-nation state conflict, global health, international crime, supply chain and terrorism. Explore Security and Resiliency
The OODA community includes a broad group of decision-makers, analysts, entrepreneurs, government leaders and tech creators. Interact with and learn from your peers via online monthly meetings, OODA Salons, the OODAcast, in-person conferences and an online forum. For the most sensitive discussions interact with executive leaders via a closed Wickr channel. The community also has access to a member only video library. Explore The OODA Community