Navigating the AI Journey: Implications and Considerations
The adoption of Artificial Intelligence (AI) is no longer a futuristic concept; it is a current reality that is reshaping industries and creating new avenues for innovation and growth.
Yet, the AI path is riddled with complexities and strategic decisions. Often, Organisations start on an AI journey that can be derailed by unforeseen surprises or the realisation that AI is not the answer to every single business problem.
Kavita Ganesan, the author of “The Business Case for AI”, goes as far as suggesting that AI initiatives should start with “no AI at all”. This approach may allow you to understand or define the problem better, ensure the necessary data is in place, and also ensure that the organisation is ready to manage the change.
So, what are the things your Organisation needs to consider before embarking on an AI journey?
1. Ethical Considerations
AI solutions, by their very nature, learn to make decisions based on the data they are fed. With this comes the inherent risk of reflecting and perpetuating any existing biases within the training data. Organisations must ensure that their AI solutions are ethically designed and trained.
There are many (now infamous) case studies where AI projects raised ethical concerns, for example:
· Amazon’s AI-powered recruiting tool that preferred male candidates over female candidates for technical jobs.
· Microsoft’s chatbot named Tay, which was designed to learn from interactions with users, started posting offensive and inflammatory tweets within 24 hours after being released.
· Misclassification of people based on racial bias by Google Photos.
Many other case studies illustrate the complexity of ethical challenges in AI. The lesson is that Organisations should pay attention to the potential impacts of AI systems on society and the importance of incorporating ethical principles into the development of AI technologies.
2. Compliance Elements
AI applications can significantly enhance decision-making and operational efficiency with their capacity to process and analyse vast quantities of information. However, this often involves handling sensitive personal data, which can raise substantial privacy concerns and necessitate strict adherence to data protection laws.
To ensure compliance and maintain public trust, Organisations should align their data handling practices with the relevant laws and regulations. Other options to consider include collecting only the data that is absolutely necessary and employing strategies such as consent management and robust encryption.
B.F. Skinner (Psychologist, Author and one of the pioneers of modern behaviourism) made the following statement in 1969:
“The real problem is not whether machines think, but whether men do.”
This quote underscores the importance of human decision-making in the design and development of AI and stresses that the onus is on humans to ensure AI complies with legal and ethical standards. It is not enough for machines to merely process data; humans must think critically about how this data is used and protected.
3. Skills Gap
The successful deployment and integration of AI technologies into business processes hinge on the availability of a specialised skill set that is often not readily available within many Organisations. This skills gap can pose a significant barrier to leveraging the full potential of AI. Organisations may need to consider several strategic approaches to bridge this gap, ranging from recruitment and upskilling to partnerships and outsourcing.
4. Financial Investment
Investing in AI is no cheap round. AI projects often require a significant financial commitment that extends well beyond the technology. Other costs include data preparation, skills and training, change management, compliance and security, ongoing maintenance and scaling, licencing, integration, and deployment. First, understanding the possible business value of your AI project may guide your Organisation in deciding whether the project is worth the upfront and ongoing investment. Consider the benefits of a phased approach and start with a pilot project that requires a smaller financial commitment, learn from the pilot and gain a clearer understanding of both the business value and the extent of the financial implications.
An excellent example of where the investment did not justify the benefit is the case of IBM’s Watson for Oncology, where the company channelled billions into harnessing advanced AI to distil insights from clinical documentation, aiming to identify the most effective treatment for cancer patients. Despite the substantial investment, the initiative failed to achieve the expected market penetration and impact.
5. Cultural Shift and Change Management
Adopting AI technologies often calls for a transformation within an Organisation’s culture and way of work. This shift requires people to adapt to new operational paradigms and embrace a mindset that welcomes continuous learning and innovation. Resistance to change is a natural human response, and it can be one of the most significant barriers to successful AI integration. Organisations should incorporate a robust change management plan as part of the AI implementation to navigate this transition. Communicate the benefits and need for AI adoption, ensure leaders are visibly engaged and championing the change, foster inclusivity and provide training and support to people to smooth the transition, reduce resistance and build a more innovative and resilient future.
6. Risk of Failure
Like any technological investment, there is a risk that the AI project will not deliver the expected results. The path to AI maturity is often non-linear and may have many unexpected challenges. Adopting a risk-aware mindset and understanding that risk is an integral part of innovation allows Organisations to plan for these derailments. Other ways in which you can prepare for these challenges include implementing agile methodologies, fostering a culture of experimentation, and evolving as you learn along the way.
7. Intellectual Property Concerns
When it comes to AI, intellectual property (IP) concerns are multifaceted. AI can create new IP, such as algorithms, models, and even content, while also being used to analyse and potentially infringe upon existing IP. Understanding the data, algorithms, and models and how to protect your Organisation’s innovations are crucial.
8. Impact on Customer Experience
As we all know (and dream), AI has the potential to revolutionise customer experience by providing personalised, efficient, and engaging interactions. One of the best examples is Amazon’s use of an AI recommendation engine. It analyses your past browsing and purchase history to suggest other products that may interest you, tailor the search results it displays and even adjust pricing based on competitor pricing, supply and demand and your buying patterns. This recommendation engine is a core part of Amazon’s strategy to enhance customer experience and increase purchase frequency and volume.
AI can do amazing things for your Organisation’s Customer Experience. Keep in mind, however, that you need to consider the privacy and trust implications of such advancements.
From the list mentioned above, it is clear that the AI journey is laden with considerations that extend beyond the initial excitement of technological innovation and require strategic foresight and a nuanced understanding of the broader business ecosystem. The key to this journey is also to look at long-term sustainability. But how do you ensure the long-term sustainability of AI projects?
Apriorit (June 2022) lists six traits of a strong AI project. The article stresses that the sustainability of AI projects is not just a technical challenge but also a strategic business consideration, requiring careful planning, a clear understanding of the value proposition and a commitment to long-term efficiency and adaptability. Here are the traits of a sustainable AI project implementation:
· Implementability: The feasibility of an AI project hinges on the availability of technology, quality data, and skilled professionals. Projects that lack in any of these areas face increased risks and potential failure.
· Viability: AI initiatives must be resource-efficient and have a clear path to profitability, either by generating sufficient early profit or having the funding to sustain operations until profitability is achieved.
· Value Creation: A successful AI project should reduce costs, increase revenue, enable new business ventures, or enhance customer experience. The specific business or technical objectives must be clearly defined to assess the potential value brought by AI.
· Efficiency: Prioritise AI tasks that can be completed swiftly and are likely to succeed. Early wins can validate AI’s effectiveness and secure further investment and stakeholder buy-in.
· Market Fit: AI solutions must be tailored to the needs of the target audience, industry, and regulatory environment. This ensures relevance and competitiveness within a specific niche, making it easier to garner stakeholder and investor support.
· Long-term Cost-Efficiency: Sustainable AI projects should contribute to ongoing cost optimisation by automating processes, providing data-driven forecasts, and reducing human error, thereby maintaining cost-efficiency over time.
The AI journey is as complex as it is promising, demanding a careful balance between innovation and pragmatism. As Organisations navigate this path, they must remain cognisant of the ethical, compliance and cultural challenges that accompany the adoption of such transformative technology.
The lessons learned from past projects serve as cautionary tales for future initiatives. By embracing a holistic approach that considers not only the technological capabilities but also the strategic business impact, Organisations can carve a sustainable AI path. This journey requires a commitment to continuous learning, adaptability, and a willingness to embrace new operational paradigms.
What other considerations should Organisations make when embarking on an AI journey?
Resources
Following the AI Roadmap | Accenture
AI educator Kavita Ganesan on uncovering AI opportunities in your business — The Intercom Blog
How to Start a New AI Project: 12 Questions to Ask Before Starting | Apriorit
Business Considerations Before Implementing AI | Technology Solutions | CompTIA
Where did IBM go wrong with Watson Health? (qz.com)
Learning from Tay’s Encounter with Social Media — Markkula Center for Applied Ethics (scu.edu)
The Ethical Implications of Artificial Intelligence — Microsoft 365
B. F. Skinner — Oxford Reference
Pixabay.com
About the Author:
Joanita Radivoev is a Scrum Master at Version 1.
