Principles of Ethical Framework for AI

There are following five principles:-

 Non-maleficence
o This principle deals with ensuring that the AI or any of its decisions or predictions
will not harm the human beings.
 Accountability
o It deals with setting the accountability of a person or group of people for any
misdeed of AI technologies. It can be the business, developers, or the trainers o Al
model
 Transparency
o It deals with making the decision making of AI transparent s SO that each step that
leads AI towards particular decision-making can be explained perfectly.
 Justice and Fairness
o This principle deals with quotability. It ensures that all the people receive the same
treatment from AI technology irrespective of their gender, race, colour, geographical
location, etc
 Privacy and Security
o This principle ensures that the AI technology will respect the human rights of people
and their data privacy and security are maintained. If followed well, all these
principles will lead to general acceptance by the public and establish trust in the
ethical AI framework.

Challenges in AI Ethical Framework

 Governance
o This is a challenge in which the accountability related to the poor or harmful
decisions of AI is difficult to set. AI development involves complex situations that
involve many stakeholders and development teams. Holding a Person or a group
accountable for misleading outcomes of AI is difficult.
 Explainability
o No matter what framework is chosen for ethics, sometimes it becomes difficult to
explain how AI reached a Particular outcome. This black-box approach often hinders
the transparency of ethical framework for AI development.
 Traceability
o Traceability of the originator of any particular data is also challenge. To find the
original source of any information on the basis of which training of AI model or
decision-making of Al model is done is often difficult t to trace. This becomes even
more difficult to identify if the dataset is prepared through multiple sources.
 Legal Compliance
o Ethical framework and its indicators choice as well ls their implementation can
sometimes lead to a breach of copyrights. These copyrights are part of someone's
intellectual property. When this copyrighted data is used to train the Al model
without permissions then it leads to legal concerns.
 o

Need of an Ethical framework in AI

 Fairness
o All problems solving done by AI or predictions made by AI model must be fair and
free from a any kind of discrimination. This has to be one of the objectives of AI
development An ethical framework marks sure that all the decisions made by it is
fair in all possible manner.
 Bias Prevention
o With the: application of ethical framework, bias in data or prediction algorithm can
be identified and mitigated or prevented. By preventing bias, AI predictions will not
result in discriminatory outcomes. There will be no discrimination based on colour,
race, geographical location, caste, gender, etc.
 Accountability
o The ethical framework ensures accountability for the actions of AI. If the predictions
or problem- solving proves to be harmful to someone Or against societal or legal
norms, AI accountability will determine who i is responsible for the action. AI
accountability can be assigned to the development team, owner organization or the
deployment team.
 Privacy Protection
o The training, testing and working of an Al model in real-world scenarios requires
handling wich a lare amount of data. This data may include sensitive information
that the AI system must not disclose to general public. When incorporated in the
development of AI, the cthical framework ensures that the data is handled
responsibly and individual consent must be taken before using personal data
 Transparency
o In the absence of an ethical framework, Al works as a black box. With an ethical
framework in place the decision-making process of Al is completely explained to the
general public. All the decisions that Al takes must be justified and explained . he
reasoning explains why and how AI makes a decision or make a certain prediction.
 Safety
o The operation of AI must be conducted in a manner that does not harm user or the
general Public. The ethical framework makes the AI model safe for use. It reduces
the potential to use AI technology to harm to anyone i in any capacity. It prevents AI
technology from being used as a weapon Oran autonomous surveillance machine.
 Human Autonomy
o AI is critical fields like healthcare or finance can make decisions that hinder the
autonomy of human being For instance, if a person in finance wants to earn huge
profit with moderate risk and does nor want to invest in cryptocurrency then Al
should not force or recommend investment cryptocurrency Therefore, human
autonomy decision-making must be preserved. The ethical framework ensures that
AI technology never interferes human autonomy.
 Social and legal compliance
o The ethical framework ensures that the decisions and predictions of AI technology
align with the s social and socio-economic background of its user. Developed AI
technology must also comply with Government or state regulations. All it is done if
ethical framework S in place.