ETHICAL AI
AI ETHICS: Study of morals and ethical considerations involved in the development or usage of Artificial Intelligence. As an emerging discipline, the definition of “good” or “bad” behavior for AI algorithms is a little unclear.
Challenges
- Is the new AI product harmful to any section of society or threatening human rights? (e.g.: Fear that generative AI will take away employment)
- What is our responsibility as designers of AI?
- What is the responsibility of users of AI?
- What are the regulations governing human-machine interaction? (Human-Robots morality)
Related Article::Understanding AI, ML, NLP, Neural Network and other terms
Related Article:: Immersive Technologies | Difference between (AR) Augmented Reality and (VR) Virtual Reality?
HUMAN-ROBOT MORALITY
Robots are slowly becoming an integral part of our lives. We know that robots will have to adapt to the individual to make sure that they assist their owner morally to respect his emotions, beliefs, and dignity but what we don’t know is the level of impact these robots will have on humans – ethically, physically as well as emotionally.
Challenges
Isaac Asimov introduced ‘3 Laws of Robotics‘ in his work called I, Robot:
- The robot should not injure any human
- Must obey orders given by humans except where such orders conflict with the first rule
- A robotic robot must protect its own existence as long as it doesn’t conflict with first and second laws
Ultimately, we’ll have to question the position of robots as moral agents and that will give rise to more sets of questions like – Should we punish the robots in the same way as humans when they are responsible for some bad actions? Then, why should we hold the robot responsible for bad actions when we can just switch it off?
Robots as moral agents?
To evaluate the moral status of robots,we must answer couple of things:
- Is the robot autonomous? If that autonomy is able to fulfill the goal and tasks and is not under the direct control of any other agents, then we say that the robot has effective autonomy.
- Does the robot behave intentionally? whether it is able to take complex decisions and differentiate between morally beneficial or dangerous things.
- Is the robot irresponsible? If the robot is able to perform social role, take responsibility, take care for its patients, then we can safely call it as a moral agent
INITIATIVES
These questions on ethics in the emerging technology of AI and robotics are nothing new. When the first robot was made we feared that robots would take over the world and result in the end of the human race (lots of Sci-Fi movies made that clear!). Then with the advent of ChatGPT and other generative AI, we pressed on the issue of unemployment.
These things have been a concern for stakeholders for a long time, the good news is we’ve begun to implement technical and legal solutions to counteract the risks:-
- UNESCO developed the 1st-ever global standard on AI ethics – ‘Recommendation on the Ethics of AI’ in 2021 and proposed a human rights-centered approach to AI. (Adopted by all 193 Member States)
- For implementation purposes, UNESCO will develop 2 practical methodologies: Ethical Impact Assessment (EIA) and RAM Readiness Assessment Methodology
- UNESCO and NIELIT establish a collaborative approach for the development of AI with ethical principles
- Many tech companies are coming up with their Ethical AI Policies and following best practices. E.g.: IBM, Apple, TCS, Deloitte, Accenture etc.
- The proposed Digital India Act, 2023, underscores the significance of addressing algorithmic biases, and copyright concerns, regulating high-risk AI systems, and promoting ethical practices while setting specific guidelines for AI intermediaries.
HUMAN RIGHTS CENTERED APPROACH TO AI
10 Points given by UNESCO:-
- Proportionality and Do No Harm: use of AI must not go beyond what is necessary; Risk assessment
- Safety and Security: Unwanted harm (safety risks), as well as vulnerabilities to attack (security risks), be avoided
- Right to Privacy and Data Protection
- Multi-stakeholder Collab and Adaptive Governance: Respect international law & national sovereignty in the use of data
- Responsibility and Accountability: AI systems should be auditable and traceable. Regular oversight, impact assessment, audit, and due diligence mechanisms in place to avoid conflicts with human rights norms and threats to environmental wellbeing
- Transparency: The ethical deployment of AI systems depends on their transparency & explainability (T&E)
- Human Oversight and Determination: Member States should ensure that AI systems do not displace ultimate human responsibility and accountability
- Sustainability: AI technologies should be assessed against their impacts on ‘sustainability’, understood as a set of constantly evolving goals including those set out in the UN’s Sustainable Development Goals.
- Awareness: Public understanding of AI, accessible education, civic engagement, AI ethics training, media & information literacy
- Non-Discrimination: AI actors should promote social justice, fairness, and non-discrimination while taking an inclusive approach
WAY FORWARD
- A comprehensive Code of Ethics for AI and Human-Robot Interaction with legal backing is needed.
- AI awareness platform
- Need for a regulating body
- Ensure regular ethical auditing, risk and impact assessment
- Develop technical expertise: Provide training to professionals in cyber security, AI/ML, ethical AI, data privacy protection, etc.