How to Develop and Adopt Ethical Artificial Intelligence (AI) Technologies in Projects

 

With the significant explosion and leap from Digital to AI technologies, organisations are often ‘out at sea’ when developing and adopting ethical AI technologies. This responsibility is typically owned by Project Management and Transformation professionals, as strategies are developed and executed.

Whilst the positive and negative benefits of AI solutions are still being debated and realised; a greater challenge exists, in that there is not a wide-spread adoption of global Artificial Intelligence (AI) standards and regulations. Whilst standards do exist in most countries, these differ substantially from country to country, making it a great challenge to adopt a universal approach to ethical AI for employees and customers.

A recent interview with Giovanni Leoni, a global industry leader and Advisory Board member for EAIGG: Ethical AI Governance Group, provides valuable advice – that the presence of global standards (ethical technical and process guidance and law) will become widespread and start to normalise during the years of 2023 to 2025.     

At the moment, the US are in the forefront of standardisation; with the release of the AI Risk Management Framework; whilst the EU are leading in terms of regulation (EU AI Act); therefore both are influencing each other.

Leoni suggests, ‘In the absence of fully adopted standards, it is advised that organisations should start by aligning their AI practices with their organisational ethical policy / code.’  This includes the organisation’s training and ethical guardrails of all operational and project activities. Therefore today, there is an opportunity for organisations to initiate proactive work and commence the change management journey, which takes multiple years.

In addition to the adoption of an organisation ethical policy / code, PMI’s Code of Ethics & Professional Conduct and other useful resources should also be adopted.     https://www.pmi.org/codeofethics

The three most common considerations of ethical AI are: how can the best data be provided; what data oversight is provided to inform effective decision making by the C-suite and Boards; and what are the ramifications of the new AI technologies?

Leoni goes on to provide solid recommendations and considerations to Transformation and Project Professionals, based on common challenges experienced across global organisations to date:

1. Project Management AI Subject Matter expertise and upskilling

What is the competency and skill set required from Transformation and Project Managers?  This can be extended to Business Analysts, Data professionals and Change Managers.  What AI capabilities are required and what upskilling and guidance are available?

2. Risk Management

As organisations embark on AI strategy and initiatives, it is fundamental to implement Risk governance - providing decision making around the risk appetite of the AI solution; the associated ethical dilemmas and how these risks will be managed and mitigated.

3. Navigating Complexity

To cater for the complexity of AI ethical implications and solutions, implementing a staged / decision gate process (Go / No Go) will allow a hard line of decision making against deliverables around functional requirements and ethical applicability.

4. Importance of Data

Understanding the dynamic of data shifting over time regarding data engineering, data science, data quality and integrity, data demand, the level of control and organisational maintenance is imperative.

5. Change Management Approach

When mapping the change impact of AI on customers and employees, it is imperative that the ‘power’ is given back to individuals; by ensuring that everyone is well informed of the AI approach and that the use of data is transparent. Some areas more likely to be of concern are building trust, removing biases, ownership of data, moral approaches, loss of control, consent and discrimination.

The level to which the detail of data is shared should always benefit the individual first and then the organisation; with reference to productivity efficiency. An example of this is monitoring the keystrokes of individuals. This is to avoid insecurity and psycho-socially negative environments.

6. Procurement

When sourcing AI solutions, partners and consultants, it is important that they align and trainon the organisation’s ethical policy.

7. Benefits Realisation

The value and benefits realised from AI need to benefit planet, people and profit.

Two key questions need to be asked:

From a business perspective – will the solution deliver value?

From a technology perspective, will the technology work?

In conclusion, I look forward to your insight as to how you have navigated ethical AI in your organisation, with the best interest of your customers and your employees in mind.