**5. Guidelines and legislations**

Many countries have published AI policy guidelines. These guidelines provide a broad level objective for the use of AI – to ensure human-centric, safe, and trustworthy AI. Most guidelines make the organization using AI responsible and accountable for their decisions and ask for the same ethical standards in AI-driven decisions as in human-driven decisions.

General key points achieved from global guidelines are:


Specially for virtual assistants, as defined above, five emerging ethical principles that are deemed important – Transparency, Justice & fairness, Non-maleficence, Responsibly and Privacy.

Legislation has been passed in California [36] to ensure Transparency of Virtual assistants. This law makes it mandatory for Virtual assistants (Bots) to disclose that they are not a real person and are virtual. Many other countries are passing laws and issuing guidelines to make it mandatory for Designers and developers to develop ethical Virtual assistants.

Many commercial organization have also issued their ethical guidelines. IBM has established an Ethics Board [37] which provides governance, review and decision making processes for IBM on ethics policies, practices, communication, research, products and services. They have also published open source toolkits which designers and developers can use to test whether there machine learning, AI models are transparent, fair and explainable.

#### *Virtual Assistant*

Google Deepmind [38] has established a focus group which focuses on ethical standards and safety. They look it from the lens of Privacy, transparency and fairness; AI morality and values; Governance and accountability; AI & worlds complex challenges, Misuse and unintended consequences; Economic impact and inclusion.

Microsoft [39] has issued guidelines for responsible bots, which are aimed at helping designers and developers to design a bot that builds trust in the company and service that the bot represents.

Many other companies have issued guidelines to ensure that the Virtual assistants developed on their platform maintain high ethical standards [40] like use supervised learning, divert issues on serios topics, do not spam users, keep user privacy, no advertisements and so on.

#### **6. What comes next**

Virtual Assistant AI technology is growing at exponential pace. In the next few years we will have virtual assistants that surpass an average humans ability to respond sensibly and specially to a consumers question. Nick Bostrom [41] presents an interesting perspective on super intelligent moral thinking. In the distant future, as AI capabilities surpasses human intelligence, it could do better than human thinkers and have the correct answers on ethics by weighing up evidence. We have already started seeing the initial versions of these intelligent machines.

IBM Debater [42] is an example of super intelligent system. This AI system can independently debate a human and provide persuasive arguments on complex topics. The system is able to listen and understand a long spontaneous speech, model human dilemmas to form an argument and generate a whole speech of an opinion and deliver is persuasively. The system has participated live and won many debate competitions.

Another example is from Soul Machines [43]. It provides Digital people ie animation of life like people on the screen. These screen animations of people is similar to an actual human who speak with expressions (eye, lips and facial movements). This provides a comfort feeling when interreacting with virtual assistant.

As virtual assistants become a part of our daily life, ethical issues regarding virtual assistants will continue to grow. It is important for the society at large to discuss and agree on the ethical principles of Transparency, Justice & fairness, Nonmaleficence, Responsibly and Privacy for virtual assistants.
