Who is responsible for AI and its fallout as the technology improves?

The era of Gen-AI may be well underway, but the job of regulating AI is still in its infancy. With so many different jurisdictions, frameworks, and forecasts about what AI may soon be capable of, it’s hard to see a cohesive governance strategy emerging anytime soon.

However, just because a global regulatory framework for AI may seem far-fetched, business leaders are not off the hook. Smart, forward-thinking leaders need to consider how to take responsibility for AI outcomes across their own organisations. It’s neither ethical nor responsible to benefit from the efficiency gains of AI without also taking responsibility for its potential consequences.

However, this responsibility may look different depending on the context. It may be convincing boards and investors that governance is a long-term investment that requires resource allocation within an organisation. Or, it may require building custom frameworks. Whatever it looks like, the accountability for how a company uses AI lies with its leaders.

Let’s take a look at some of the key considerations business leaders need to make as the technology evolves.

A shared responsibility

As our EVP & Partner Solange Sobral recently pointed out, recent efforts in AI regulation — such as the EU’s AI act and similar ongoing discussions in the UK — “show how tough it is to create a cohesive global regulatory environment. This gap leaves organisations in a bit of a governance limbo.”

Adding to this is the fact that AI development and deployment often involve a chain of actors, making singular blame difficult. A common response from business leaders to all that complexity is to simply say, “We’ll wait and see what happens with AI regulations from governments.” But that approach cannot hold.

Indeed, the first step in approaching this problem is a paradigm shift in how we think about it. Deploying AI responsibly as its capabilities grow is not someone else’s problem to figure out. It’s everyone’s problem — from governments and corporations to engineers and leaders of AI companies themselves. Changing the culture to one of shared responsibility is essential to moving forward

What leaders can do

There’s a reason why an international regulatory framework for AI is hard to come by. Unlike traditional software, AI's adaptive and learning nature makes it harder to pre-determine outcomes or trace errors. Plus, the black box nature of AI can make it hard to understand or determine how complex AI models arrive at their decisions.

But there are things business leaders can do.

The first is to keep the human power of discernment at the heart of an organisation’s AI strategy. Business leaders must ensure that AI never replaces the ethical reasoning, contextual awareness, and empathy that only people can provide. This means designing workflows where humans are involved in interpreting AI outputs, especially for critical decisions that affect people’s rights, safety, or livelihoods. A culture of overreliance on AI should be avoided at all costs, even as AI gets more and more sophisticated.

The second is to shift the focus from finding fault to risk management. Rather than looking to assign blame when things go wrong, leaders should be proactively safeguarding against poor outcomes before things happen. There are a host of ways that AI might go wrong, be it from unintended bias, misuse, or algorithmic drift. In the absence of a universal regulatory framework, internal risk management becomes a key priority. Leaders must prioritise scenario planning, impact assessments, and continuous monitoring to identify and mitigate potential risks early.

Next, leaders need to ensure they are using high-quality and unbiased data to inform their AI systems. AI outcomes can only be as good as the data on which they are built. This requires a conscious effort to audit data sources, involve diverse stakeholders in data collection, and test models for fairness and blind spots. By prioritising data quality and inclusivity, leaders can help prevent harmful or discriminatory outcomes and build trust in their AI solutions.

Finally, training and ethical guidelines need to be viewed as an ongoing project. Unleashing AI on your organisation and hoping everyone catches up isn’t sufficient. Leaders need to make AI knowledge and literacy accessible across all levels of the organisation. This means investing in training so employees understand how AI works, where it can fail, and what to do when it does. Leaders should establish clear ethical guidelines for AI use and create robust channels to report concerns or escalate issues. This democratisation ensures that accountability is not isolated at the top, but shared throughout the organisation

In all of these steps, it’s crucial for leaders to strike a balance between innovation and protection. Regulations and governance isn’t about diminishing or inhibiting the potential of AI. They are about enjoying the many efficiency gains AI offers while still keeping ethical outcomes as a primary goal.

---