A Risk Management Approach to Responsible AI Adoption 

Artificial Intelligence is here to stay. But that doesn’t mean we should throw up our hands and let AI run roughshod over otherwise sound product development operations, right? The stakes are too high, and much of the technology’s potential (for good and for ill) remains unknown. 

Since our founding nearly 3 decades ago, ITX has been committed to helping clients manage risk. In that time, we’ve learned that diving headlong into any new technology – let alone one as potentially disruptive as AI – without careful exploration invites all sorts of risks: financial, reputational, ethical, and legal. 

We documented our approach to AI integration in a position paper, ITX’s Perspective on AI and Our Approach to Adoption. It reflects a clear statement of our current thinking – as AI evolves, so too will our assessment – complete with practical governance framework and set of rules that guide our behavior. This blog describes that framework and explains what AI governance means inside ITX. 

Why Governance Matters 

Not wanting to be left behind, many companies charged impulsively into AI adoption, often with dire consequences.  

In August 2023, Gannett shut down the use of an AI tool meant to help it produce more news posts faster – but only after publishing articles of embarrassingly mediocre quality emerged. In an apparent race to publish chat-enabled Q&A tools, both H&R Block and TurboTax released chat bots in advance of the 2024 tax season. By March of 2024, both were lambasted by the Washington Post for delivering wrong answers upwards of 50% of the time (account required to view).  

These real-life scenarios aren’t just the result of a rush to deliver tools using AI. Encouraged by end users and eager business leaders, some software companies are out in the market pushing their clients toward “solutions” just like this – regardless of the consequences. 

Effective Leadership in the Age of AI 

Leading in the AI era isn’t about always being first. Instead, effective leadership means being methodical, purposeful – even deliberate – taking the time to learn and grow, proceeding with caution and care.  

This new phase of digital transformation calls on business leaders to secure a solid grasp of AI’s rewards and risks. A trustworthy technology partner can help, someone who understands your big-picture business objectives and embraces them as their own. 

As companies blend AI into their product strategy, effective leadership means establishing a framework for governing the integration of AI tools by acting as a steward for responsible, ethical AI implementation and adoption.  

Governing AI at ITX 

It’s important to remember that as a human-created technology, AI is susceptible to human biases and errors that can result in discrimination and other harm to others. AI governance addresses the inherent flaws arising from the human element in AI creation, application, and maintenance.  

When we talk about AI governance at ITX, we’re referring to the processes, standards, and guardrails that help ensure AI systems and tools are safe and ethical. We see our responsibility for governance in terms of a structured approach to monitor, mitigate, and confront potential risks while supporting innovation and building client trust.  

A Principled Approach for Integrating AI at ITX 

AI governance starts with a principled approach: we believe that AI helps us be better product managers, architects, researchers, designers, engineers, and QA specialists. ITX’s integration of AI tools leans on four key principles: 

• Empowering People. AI amplifies human abilities by reducing time spent on repetitive tasks and freeing team members for creative problem-solving and high-value work.  

• Responsible Governance. Transparency, security, risk mitigation, and ethics guide our AI implementation. 

• Growing to Perform More Powerfully. We embrace AI integration as a journey, continuously adapting and upskilling our team to thrive in this evolving landscape.  

• Reducing Cost, Improving Quality. AI-enabled workflows show promise when it comes to reducing cost and/or improving quality, and we’re eager to discover both for our clients. 

From these principles we have established a set of 10 rules that serve as behavioral guardrails – i.e., constraints that monitor our “inherent human flaws” and that drive creativity and innovation.  

Learn more about ITX’s values-driven approach to AI and the Rules that guide our interaction with it at the ITX Artificial Intelligence Hub. 

10 Rules Governing AI at ITX 

We strive to deliver to clients all the benefits that AI has to offer while simultaneously limiting their exposure. To achieve this objective, we’ve carved out a set of rules that govern ITX’s approach to, and application of, AI tools. 

1. Values driven, first and always. ITX will use AI in ways that  are consistent with our Corporate Values.  

2. For humans, by humans, with humans. A human decisionmaker will remain in control when integrating AI into ITX workflows.  

3. We drive efficiency and quality to create client value. ITX embraces the growth  mindset; thus, we will explore any tool that delivers value to our clients and efficiency to our teams.  

4. We abide the law. Without exception, ITX respects the laws and regulations regarding the use of AI within every jurisdiction in which we operate.  

5. Our clients determine their own risk tolerance. We prioritize privacy, indemnification, and confidentiality; we advise clients on risk-benefit tradeoffs and respect our clients’ risk tolerances; we follow their direction, and we support their decisions regarding new tools or technologies.  

6. We respect intellectual property rights. We will work to maintain confidence that we are not violating anyone’s intellectual property rights.  

7. We strive to avoid unintended consequences. We learn by performing structured experiments in controlled environments and demonstrate client value by sharing our newfound knowledge.  

8. We create client value through the technical disciplines we practice. ITX Practice Leaders uphold these rules and understand the application of AI within their respective disciplines.  

9. We innovate and manage risk through the technical disciplines we practice. Our Practice Leaders pay attention to the industry and actively seek out best practices that account for risks and opportunities.  TX Practice Leaders uphold these rules and understand the application of AI within their respective disciplines.  

10. We understand and apply these AI rules. Team members using AI will be trained on these rules and instructed to ask for guidance in any situation that is unclear.  

Examples of AI Governance In Action  

ITX’s perspective on AI governance is founded on corporate values that drive all behaviors and decisionmaking. It is nurtured by a companywide awareness and engagement that ensures AI-related systems are developed and applied with transparency and explainability in alignment with our clients’ business objectives. 

Through a handful of internal company policies our approach to AI is made clear and accessible to the entire organization. Also in those policy documents – annual review/revision of which is required – our rationale is spelled out. Established well before AI burst onto the scene, these policies are as relevant today as when they were first drafted. For example: 

• All ITX-built products must adhere to internal GIT standards.  

• Use of source code generated by an LLM (e.g., ChatGPT) requires approval from ITX General Counsel before deployment. Approval is based on a fully executed client agreement permitting such use.  

• ITX service agreements now memorialize the client’s commitment with language that clearly authorizes our use of AI tools during the engagement. 

• Decisions made by automated AI tools are not applied without human guidance; an ITX decisionmaker remains in control to make sure that decisions align with our clients’ objectives. 

• Practice leaders across ITX technical practice areas are accountable for understanding our 10 Rules and for developing the team member skills sets and industry-standard best practices. 

External Support from Corporate Partners   

We understand the legal, financial, and reputational risks our clients face; no AI governance program can eliminate all risk. 

That’s why ITX has partnered with IBM® to provide additional governance services powered by watsonx™. Having achieved Silver Business Partner status with IBM, ITX is able to help clients deploy watsonx to support not just governance, but also to train, validate, tune, and deploy their own AI models. 

ITX also leverages Microsoft™ Copilot as a general solution to meet many of our internal AI needs. We believe Copilot represents a high-quality solution with which to conduct our exploration into AI use cases. Like IBM, Microsoft has committed to indemnifying companies using Copilot against IP infringement claims related to the use of outputs produced by Copilot. 

---