Ethical AI Leader Roles: Realities for Employers & Candidates

The world of AI technology is really flourishing right now. But so are AI fears.

Organisations of all kinds are concerned about the murky regulatory waters that currently surround AI. They’re concerned about maximising profits and minimising unforeseen risk. Many are rightly concerned about doing the right thing for their customers and customer data.

Yet despite these fears, AI uptake ploughs on unabated. Workplaces are using AI to enhance workflows in numerous ways, whether they have responsible AI guidance in their midst or not.

This landscape makes hiring a responsible AI leader unavoidable to many companies.

So whether you’re looking to hire an AI ethics lead in order to de-risk or you’re exploring a career move towards AI ethics leadership, join us as we peel back the layers of this complex, vital, and growing area of business.

9 Truths About Ethical
AI Leadership

Employers Are Rarely AI Ethics Experts

Most employers are seeking ethical AI leadership for a reason: because they lack that particular expertise and vision. So naturally, they go looking for a trustworthy, knowledgeable party to guide them.

This puts a lot of power in the hands of candidates that come their way to fill that role, because they are instantly the most knowledgeable party in the room. The employer needs them – not the other way around!

In fact, this is one of the reasons why the team and I love working ethical AI leadership roles. A candidate can come along, assess an employer’s AI and data practices, and really take them to task about what they’re doing wrong – potentially transforming the way an organisation looks at AI and risk for the better.

Employers stand to benefit greatly from this more humble approach. They can better understand where their own weak points lie; they can better understand each candidate’s strengths and areas of focus; and in turn, they can understand which party is going to steer them in the direction they need.

AI Ethics Isn’t a Data Science Role

Anyone outside of the world of data ethics would likely be forgiven for thinking that AI ethicists would require a heavy, academic, data science background in order to do what they do. And yes, a hefty dollop of data science fundamentals certainly won’t hurt any budding AI ethics leader’s chances.

However, in all of the ethical AI roles we’ve filled thus far, the successful candidates actually had far more experience in ethical, legal, and regulatory areas of business rather than heavily technical or data-focused AI roles.

Ideal ethical AI leadership candidates are those who can untangle legal obligations, potential risks, and technical practicalities – steering the organisation towards solutions that seamlessly unite these three concerns.

(Side note: If you’re a data scientist who’s eyeing a move towards ethical AI roles, you might want to start networking in regulatory, de-risking, and legal circles.)

Ethical AI Reporting Lines Can Differ – And Usually Aren’t Technical

Unlike more established areas of business like finance and marketing, the brand new field of AI is still finding its place in the corporate org chart (cybersecurity is in a similar position).

Where an organisation places its AI and AI ethics teams can depend heavily on that organisation’s own technical, risk, and data landscape. We’ve seen AI ethics leaders reporting to a wide array of business functions, including the CEO, CTO, Chief Risk Officers, Legal Officers, and Data Officers.

According to IBM, between 2018 and 2021, responsibility for AI ethics shifted significantly from technical to non-technical leaders. Though this is before the AI floodgates were opened by the likes of ChatGPT in late 2022, it still shows a marked – and quick – maturation of the field before that time.

Keeping AI ethics apart from technical AI functions might initially seem paradoxical. But there are two main reasons why it’s great practice. The first is the difference in focus. Ethical AI leaders and teams are more focused on avoiding legal, regulatory, and moral risk to an organisation, so are naturally more at home amongst business functions focused on de-risking and legal matters.

The second is minimising conflict of interest. The role of an AI ethicist often involves policing technical AI teams to avoid a potential ethical issue. When these ethics and technical functions become more intermingled however, it creates a less objective power dynamic between the two skill sets – one that could result in the tech teams seemingly “marking their own homework” so to speak. Therefore maintaining some distinction between the two areas shows maturity towards AI and the risks it presents.

Reasons for Recruiting an AI Ethics Lead Vary Wildly

Here’s a snippet from IBM’s guide, “The CEO’s guide to generative AI: Ethics can’t be delegated” about those who got in on the AI gold rush early:

Generative AI is like the wild west. The rush to riches has outpaced rules and regulations—and early prospectors have the chance to strike it rich. But at what cost? Organizations that push forward without considering the intricacies of AI ethics and data integrity risk damaging their reputation for short-term gains.

The same guide presents a worrying statistic. Though 79% of executives say AI ethics is important to their enterprise-wide AI approach, less than 25% have implemented common AI ethics principles. It doesn’t take a de-risking genius to see how, lacking such principles, hiring an AI ethics leader may sometimes be a bit of a “panic buy” rather than a planned, strategic decision.

Yet somewhat less cynically, if an organisation’s brand reputation and core selling point hinges on being ethical, and they are starting to significantly adopt AI, then bringing on ethical AI expertise might be a no-brainer!

Project & Transformation Management Skills Are Useful

On its face, being in charge of an organisation’s ethics (AI-related or otherwise) may not initially appear particularly transformative.

Yet any kind of ethics leader will likely need to change the way things are happening in order to make the organisation more responsible. This can require human change management skills to change behaviour and culture; technical transformation skills to assist in steering AI adoption in the right direction; and project management skills to manage iterative changes towards a more risk-free future.

Therefore, AI ethics leaders may find any kind of transformation skills and experience surprisingly useful.

Corporate Stakeholder Experience is Non-Negotiable

All board-adjacent leaders need to be au fait with communicating with the C-Suite. But any AI leaders who will be communicating directly with the board may have a particularly tricky job on their hands.

Let’s start with the basics. Any leader communicating with the board will need to understand how the C-Suite works, grasp the unique politics that are often at play, and get to know each member of the board in order to communicate with them personally and effectively.

Navigating complex and nuanced ethics and risk evasion with a board can be enough of an uphill battle as it is. But doing so within the often misunderstood field of AI can be particularly challenging. AI risks are rarely black and white, and regulatory guidance and legislation around AI is still very much a work in progress, so you will need to help the board traverse these murky, uncharted waters.

And as IBM rightly points out, “Some companies freeze in the headlights of regulatory ambiguity.” Therefore, communicating AI risk to the board can require quite sophisticated levels of stakeholder interaction.

The Responsible AI Rabbit Hole Runs Surprisingly Deep

Maintaining organisational ethics, steering clear of regulatory concerns, and balancing business risk are all significant responsibilities. However, the rabbit hole that AI ethics leaders have to traverse actually runs a lot deeper.

Here are some of the AI-related ethical concerns that we’ve discussed with candidates and partners in the past – many of which lie far beyond merely “using data and technology responsibly and legally”:

● How did we source the data our tech solutions use in the first place? Was it obtained ethically?
● How do we work alongside AI without asking it to do all thinking and creativity for us?
● How do we make the most of AI’s potential in areas of the business that require a human touch?
● How do we use AI in a way that has minimal negative and maximum positive societal impact?
● How do we fulfil our carbon-reduction commitments while using AI, given that it is orders of magnitude more energy-guzzling than conventional computing?
● How can we anonymise the data we feed into AI tools to eradicate bias, especially around protected characteristics?
● How can we ensure that the AI tools don’t create or introduce new biases via machine learning inference?

Taking a complete, top-down view of an organisation’s AI ethics like this might make some people’s heads spin! But some of my fellow armchair philosophers might consider it a really juicy challenge.

We Need AI Transparency More Than Ever

Transparency in data and data processing is an important ethical concern with AI; especially for those of us who are in jurisdictions that use legislation like GDPR.

Yet many AI and machine learning tools, especially larger language models, suffer from what is called the “black box” problem: where AI models are so complex, with countless data points and inferences, that not even their creators can state why they provide certain answers and come to certain conclusions.

Even in 2021, before the current AI boom where everyone and their dog is using AI in some form, FICO reported that 65% of respondents’ companies can’t explain how specific AI model decisions or predictions are made, and only a fifth monitored their AI models for fairness and ethics. We’re sure those figures are different now, but still – yikes.

Any organisation’s issue of AI transparency isn’t going to be solved by one person in a relatively casual afternoon. But in-house ethicists can try to unpick how existing AI tools operate and seek to embed some level of transparency into new AI projects.

Ethics Could (and Should) Become the Most Important Role in AI

AI is helping organisations unlock the latent power of data and automation to immense benefit. Yet with great data power comes great data responsibility!

As AI continues to assert its presence in the public zeitgeist, most of us have been party to conversations involving AI’s ethics. Whether it’s pondering AI’s potential to replace entry level workers, discussing the fears of deepfake media, concerns around use of other people’s intellectual property to train AI models, or the increasing lurch towards autonomous technologies and decision-making.

We would argue that, in this current climate, any organisation using AI needs some grounding in AI ethics in order to operate. But as AI technology matures, the need for AI ethics within any organisation will become unavoidably essential.

I’ll let the AI bros and bro-ettes argue about how far we are away from Artificial General Intelligence, but creating a digitally modelled system that matches or surpasses the capabilities of a human brain brings with it its own, much deeper set of ethics. But you can watch Blade Runner in your own time to get up to speed on that.

Either way, we’re going to need a lot more AI ethicists rather soon…

Who is Responsible for Responsible AI?

Ethics officers – AI-focused and otherwise – have an incredibly important role in any organisation.

We see arguably unethical corporate behaviour in the tech world all the time; personal data misuses, copyright infringement, cyber-insecure practices, and much more. And tools that use emerging tech like AI and blockchain are already being used for ill intent out in the wild – and in ways those tools’ creators would likely be horrified by.

It’s the job of the ethicist to be that common sense voice in the room that steers companies away from malpractice. They’re not just there to keep an organisation’s nose clean to legal teams and regulators – they’re there to keep an organisation accountable to doing the right thing.

But not just as a hollow PR play. In a way that benefits us all.