More regulation sought over firms’ use of AI

Capgemini Research Institute surveyed 4,400 consumers in six countries (UK, US, France, Germany, the Netherlands and China) and 1,580 professionals from 500 organisations to explore attitudes towards ethics and transparency in AI.

Of the consumers surveyed in the study, 76% think there should be further regulation on how companies use AI, while 75% agreed they want more transparency from companies if a service is based on AI.

Executives in nine out of 10 organisations said they are aware of at least one instance of AI systems resulting in ethical issues, such as collecting and processing healthcare personal data in AI algorithms without consent, AI that is not able to explain how it makes decisions on credit or insurance claims, or using AI for workplace surveillance without employees’ consent. Almost half of the consumers surveyed ( 47%) believed they had felt the impact of such an issue.

Consumers also claimed that they would place more trust in companies whose AI interactions they perceive to be ethical ( 62%). However, 41% said they would complain if they experienced an ‘unethical AI interaction’ and 34% said they would stop engaging with the company.

Over half of the professionals in the study ( 51%) agreed it’s important to ensure AI systems are ethical and transparent, and 41% claimed that their company had abandoned or would abandon a system altogether if it was found to have caused an ethical issue. Additionally, 44% of employees claimed that they or their colleagues had raised concern about potentially harmful use of AI systems within their organisation.

UK executives in the study were most confident of all the countries surveyed that their organisations’ AI systems are transparent ( 36% compared to 25% globally) but only 28% said their systems are ethical or fair.

Anne-Laure Thieullent, AI and analytics group offer leader at Capgemini, said: “This research shows that organisations must create ethical systems and practices for the use of AI if they are to gain people’s trust. This is not just a compliance issue, but one that can create a significant benefit in terms of loyalty, endorsement and engagement.

“To achieve this, organisations need to focus on putting the right governance structures in place, they must not only define a code of conduct based on their own values, but also implement it as an ‘ethics-by-design’ approach, and, above all, focus on informing and empowering people in how they interact with AI solutions.”