The Business Analytics Ethics Compass for Data and Artificial Intelligence (AI) analyzes seven dimensions that are instrumental in determining the ethical maturity of your organization:
Start the Business Analytics preview. It contains an overview of all dimensions of the analysis and offers you a first insight into the question structure at hand with one of many response scales.
1/16 Principle 1: Human Centricity
Humanity and human care Data analytics and AI solutions are designed and developed around and for the people. In particular, data and AI solutions are designed to augment, complement and empower human cognitive, social and cultural skills.
2/16 Principle 1: Human Centricity
Understanding public attitudes and ethical expectations We understand the community social attitudes, expectations and concerns in relation to trusting data analytics and AI solutions and reflect these considerations in our business practices.
3/16 Principle 2: Shared prosperity
Societal prosperity Data analytics and AI solutions are designed – where possible – to deliver benefits to the public (shared prosperity) in order to solve societal problems.
4/16 Principle 2: Shared prosperity
Monitoring practices We regularly monitor the impact of our AI/data analytics solutions on institutions, society, democracy and people to assess the benefits (as well as risks) of our data analytics and AI solutions.
5/16 Principle 3: Understanding, addressing and balancing competing needs
People’s needs and data proportionality When collecting data for the purposes of developing data analytics and AI solutions, we understand the people’s needs, and we evaluate the relevance and proportionality of the data being collected against those needs.
6/16 Principle 3: Understanding, addressing and balancing competing needs
Balancing consumer and employees’ needs We assess the impact of the data analytics and AI solutions on our employees. These impacts are taken into considerations to find the right balance between employees, consumers and business needs.
7/16 Principle 4: Human autonomy, empowerment and oversight
Human autonomy We have controls in place to prevent data and AI solutions from being used to subordinate, coerce, deceive, manipulate condition or nudge humans.
8/16 Principle 4: Human autonomy, empowerment and oversight
Human empowerment - explainability Decisions (automated, semi-automated, and human) produced or informed by AI and data analytics solutions are documented and articulated in ways that are understandable by internal (e.g. internal auditors, employees) and external stakeholders (e.g. consumers, regulators) irrespective of their technical background.
9/16 Principle 5: Do not harm (fairness, safety, cybersecurity and system performance)
Do not harm Our data analytics and AI solutions are designed and operate in a manner that respects, serves and protects humans’ physical and mental integrity, personal and cultural sense of identity, and satisfaction of their essential needs.
10/16 Principle 5: Do not harm (fairness, safety, cybersecurity and system performance)
Data and AI ethics - risk appetite We have defined the data and AI practices and use-cases which are deemed to be unethical, and therefore, prohibited. These include use-cases that represent a threat to human rights where interference with human rights is not deemed reasonable, necessary and proportionate (please refer to your local Human Rights Charter/Legislation).
11/16 Principle 6: Adopting ethical data privacy best practices
Data privacy - ethical best practices vs regulatory compliance Our data privacy practices are benchmarked against national and international best practices to address ethical expectations and concerns raised by the community, individuals and regulators. Examples of best practices include: data practices outlined in the ACCC inquiry on digital platforms; Germany Data Ethics Commission Recommendations issued in 2019; Recommendations issues by the UK Data Ethics and Innovation Commission following analysis of specific use-cases; U.S. Mind your Business Act.
12/16 Principle 6: Adopting ethical data privacy best practices
Data purpose We clearly articulate to the individual (consumer and employees) the purpose of collecting and processing data.
13/16 Principle 7: Data quality and integrity
Data quality We have procedures in place to assess the quality of the data being used to develop data analytics and AI solutions and address any data quality issues.
14/16 Principle 7: Data quality and integrity
Data anonymisation We assess and minimize the risk of re-identification.
15/16 Principle 8: Ethical data monetization
Third party information sharing with the public Consumers and employees are provided with a list of third parties we could share their data with; individuals are provided with the option to opt-out from having their de-identified data shared with specific third parties.
16/16 Principle 8: Ethical data monetization
Data sharing within the organisation Consumers and employees’ data is shared within the organisation only with authorised people who use this data solely for purposes aligned with the original consent provided by the consumer or employee.
Would you like to find out more? Talk to us about your current tasks and how this Business Analytics can support you in developing solutions for them. You can try out the Business Analytics free of charge first and see initial results straightaway. Plan your individual roadmap with KPMG Atlas!
We're looking forward to hearing from you.
Get in contact with us here to find out more about our service. We're looking forward to hearing from you to discuss your specific issues.
Try out the Business Analytics in full and free of charge now!
Have you already planned your next steps? Here you have the opportunity to perform an individual maturity level determination and receive an individual summary of the results immediately.
Off-the-shelf not for you?
Please contact our experts to create a non-binding custom-fit solution for your organization.
Would you like to find out more about our tailored service?
Benefit from our rapid and uncomplicated location determination for your current issues at any time by using our digital service.
The ethics of AI encompass all challenges arising from the use of artificial intelligence. With the number of problems significantly increasing in recent years, the ethics of AI have evolved as a new field for research and application. There are three important components to this: algorithm ethics, data ethics and moral judgment. Algorithm ethics address the ethical bases of an AI system, which need to be considered during the development of these algorithms. Data ethics involve the generation of data that is used for AI as data is the base for each AI system. Moral judgment focuses on the people who develop, use and design AI and how reflective they are in handling the data.
Profit orientation and morals often conflict with one another. Sometimes it is necessary to accept initial losses as a short-term investment in order to achieve profits over the long term. Ethical conduct is an investment in the value of trust, which pays off, for instance, through long-term cooperation. In times of continual new challenges, such as resource scarcity or threats to ecological foundations, long-term goals often lose their focus.
Where AI applications are used, consumers will increasingly question whether ethical principles are being adhered to. If this is the case, consumer acceptance will rise. In addition, consumers will be willing to publish positive reviews or to buy more products. By contrast, consumers will steer clear of producers exhibiting questionable, non-transparent approaches. Companies will also be aware of this development; they face the challenge of introducing AI systems quickly without jeopardizing the principles of ethics. According, ethical aspects already need to be considered right from the outset of development.
If you want to deal with artificial intelligence or use it, you need to be aware of the risks. Often users are concerned as large parts of controls are passed on to artificial intelligence systems and there is no person responsible in the case of damages. Further, large volumes of data are required to train algorithms, which means that data misuse cannot be ruled out. Transparency is another major point for criticism. AI decisions are generally not comprehensible and can therefore lead to wrong decisions. To counter this, researchers are already working on the issue of "explainable AI". An example that is often mentioned in this context is the "AMS algorithm" [job prospect profiling algorithm in Austria]. The job prospects of unemployed people are ranked into low, medium and high categories. Critics allege that the algorithm is not transparent, and that it is discriminatory and based on out-dated knowledge. In addition, decisions could potentially be based on outdated or erroneous information.