Artificial Intelligence (AI) does not always meet with open arms. The ipt expert Cyrill Rüttimann explains the ethical principles that are important.
Ensure transparency in your data sources. Make sure that the data comes exclusively from known sources. They may also only be used for the intended purpose.
Otherwise, not only will the trust of your customers and partners suffer, you will also violate internal departmental policies such as compliance.
«Autonomously decisive systems hold new risks. Both in terms of attack scenarios and potential manipulation,» the AI expert points out. In order to prevent such novel risks, not only a new awareness is needed, but also new processes and strategies must be developed.
It is important that AI systems are designed in a comprehensible way. In this way, errors can be isolated and corrected quickly using standardized processes.
«AI systems can make false predictions,» as Cyril Rüttimann points out. Therefore risks should be evaluated and cushioned accordingly. For example, by having the decision made by the AI system validated by a human.
Since the data for the training of the AI systems is selected by humans, bias resonates. The data may have been incomplete or unconsciously restricted.
The potential influence of bias must be considered. «When designing the system, it is important to ensure that the bias is transparent and controllable,» explains the AI expert.
Automated processes are replacing work steps that were carried out by humans. Involve the workforce and carry out the change process transparently and honestly. This promotes the acceptance of AI in the company.
Read the full article by our expert Cyrill Rüttimann published in C-Level magazine 4|2019