Using AI responsibly

Initiate ethical assessment right at the start of the procurement process when you are identifying the problems you want to solve with an AI system. Already at this stage, consider which rules and terms your organisation will apply to using AI tools.

Work with the persons responsible for the system and with senior management. The management is ultimately accountable if the system causes harm or damage.

Allow other staff members to participate as well. This may result in insights and observations that a small group of people might not have come up with.

To maximise the accountability of your practices:

1. Acknowledge and clearly state the objectives of the system and their value base.
2. Be aware of legislative considerations, including fundamental rights and international commitments.
3. Select methods for establishing data quality criteria and for ensuring quality.
4. Assess how the system can affect different groups, individuals or society in general.
5. Consult target groups and stakeholders from the start, already when setting objectives.
6. Monitor quality and impacts continuously.

Try to have empathy and put yourself in the position of the system’s target groups, and especially try to identify any risk of negative impacts. Try to genuinely cooperate with the target groups. By being proactive, you can avoid having to resolve and fix problems later.

Another challenge is the contextual nature of AI systems in relation to their ethical functioning. All data processing systems are designed in a specific time and place for a specific need, but circumstances and backgrounds become highlighted in algorithmic systems that are based on data collection and analysis.

In other words, there really is no all-encompassing, concrete template that would tell you what to do exactly.