AI Governance in Practice

Any invention has the potential to do good but, sadly, also to do bad. Artificial intelligence (AI) is no exception, seeing use as a tool for cyber security as much as it is for cyber crime, among other things. Such misuse can erode confidence in a technology that has become indispensable in everyday life, including running a business or organisation.

In fact, we’re seeing this right now with a slow pace of AI adoption among businesses and organisations. In one cybersecurity firm’s survey of more than 700 business leaders from 26 countries, most said data privacy issues and inaccurate output delay AI rollouts by up to a year. The problems persist even as they invest in employee training. (1)

If AI is to remain the norm, it’s about time we started asking ourselves the hard questions about how it should be used. Only then can businesses and organisations implement AI principles and best practices that inspire confidence and trust among the public.

To that end, the World Economic Forum’s Advancing Responsible AI Innovation playbook provides a solid roadmap for fostering AI governance. Here’s everything you need to know to use the technology responsibly.

Regardless of where you stand on AI’s inevitability, the latest numbers make it all too clear. In an OECD survey of businesses based in G-7 countries (Canada, France, Germany, Italy, Japan, the U.K., and the U.S.), over half agree that AI is “critically important.” Of these, the majority use AI for quality control, supply chain management, and logistics. (3)

There’s no denying that AI can perform automated decision-making and other tasks well. It shows attention to detail that’s measurable by its code and is far less prone to mistakes caused by fatigue. And because it doesn’t get tired, it doesn’t need paid time off.

Yet, for all its benefits, the technology is still far from perfect. Too many businesses make the mistake of leaving key tasks entirely to AI, which often doesn’t end well. At its current state, one MIT study states, it can only complete roughly a third of office tasks at most.

There’s a lot to be excited about AI’s growth, but it also pays to temper your expectations. Your business or organisational strategy should align with its innovations and milestones, lest you risk getting the wrong idea about AI.

The playbook suggests the following actions for aligning your long-term strategy: (2)

• Set responsible AI vision and principles to define your commitment
• Encourage employees to submit feedback on AI initiatives
• Enable data collaboration while limiting access to raw data (data governance)
• Keep uncertainties to a minimum by preparing contingency plans

Another reason that AI adoption flops hard among businesses and organisations is that not everyone is on board with such change. Initiatives can face resistance, especially from old hands who have been comfortable with the status quo throughout their tenure.

Some of these resistances came to a head, such as the 2023 SAG-AFTRA strike in the U.S. that caused major interruptions in the entertainment industry. In fact, that strike had only ended last July following an agreement that put AI protections for professionals in place.

The reasons for the opposition make sense if you’ve been following the news for a while. The technology has shown potential in removing humans from the equation. In the case above, professionals protested the companies’ use or plans of using generative AI to make digital media (e.g., motion capture, voice-overs) without their input.

AI adoption may be inevitable, but it doesn’t have to alienate humans. Its limitations mean it still requires humans to monitor and manage its output, more so in niches where a single inaccuracy can have grave consequences on a client or customer’s well-being. Businesses and organisations need to make their people understand that they still matter.

Key to this effort, according to the handbook, involves appointing an AI governance leader. It goes by various job titles, but this person is tasked with helping the business achieve its AI governance objectives within ethical considerations and regulatory requirements. At the core of their responsibility is an AI governance framework, which can include: (2)

• Making self-reporting AI safety risks and issues easy for employees
• Documenting every case that used or involved AI systems
• Understanding how well the company knows its AI systems
• Promoting transparency in reporting by leading by example

Augmenting activities with AI isn’t as easy as buying an AI tool or subscribing to an AI/ML model. A business or organisation is expected to use their preferred AI system long enough for changes to render its algorithm performance obsolete.

In a CXOTalk podcast last July, Lenovo Global CIO Arthur Hu talked about a trap that many businesses and organisations fall for too often: “analysis paralysis.” Exactly as it says, this is essentially overthinking to the point of being unable to come to a decision. In the case of AI adoption, Hu said it sets in because of how fast AI capabilities grow. (4)

Hu pointed out that there’s no point in fretting about AI tools and systems being outdated. Rather, he advised businesses and organisations to work with the AI they have and adapt it to changing climates as best they can. If the signs of performance degradation (e.g., using biased data) start setting in, only then you consider getting a new AI/ML model. (4)

As such, understanding an AI system’s life cycle—from acquisition to monitoring—is key in AI governance. By ensuring you can squeeze every bit of investment out of your AI tool, you can avoid frequent, costly retooling and retraining. This also benefits customers, as it cuts the risk of issues that can erode trust in the brand.

To that end, the handbook recommends the following practices: (2)

• Empowering employees to question everything involving the AI’s use
• Ensuring there are enough human staff for AI risk management
• Defining key metrics for training employees to be AI-literate

With AI being a subject of debate, using it within ethical guidelines is more important than ever. Ask our team to know how you can reap the business benefits of AI governance in every aspect, from cyber security to managed IT.

1. “Survey Finds AI Trust Eroding Amid Security Risks,” Source: https://www.securityinfowatch.com/cybersecurity/article/55322018/survey-finds-ai-trust-eroding-amid-security-risks
2. “Advancing Responsible AI Innovation: A Playbook,” Source: https://reports.weforum.org/docs/WEF_Advancing_Responsible_AI_Innovation_A_Playbook_2025.pdf
3. “The Adoption of Artificial Intelligence in Firms: New Evidence for Policymaking,” Source: https://www.oecd.org/en/publications/the-adoption-of-artificial-intelligence-in-firms_f9ef33c3-en.html
4. “AI Technology Outdated? That’s Okay, Just Keep Moving,” Source: https://www.forbes.com/sites/joemckendrick/2025/07/31/ai-technology-outdated-thats-okay-just-keep-moving/