Do We Need Ethical Considerations for AI?

Posted by Tirthankar RayChaudhuri on Dec 12,2023

There is much paranoia and pessimism about AI misconceptions being promulgated in an irresponsible manner by individuals who are ignorant about what AI/ML really is. We assert that such fear mongering is being spread using the convenience of internet media solely for the purpose of generating sensationalism of a negative kind.

There is nonetheless a real risk of AI/ML being misused for criminal and unlawful purposes. While this risk is not specific to AI alone and exists for all forms of advanced technology, the concern and risk related to AI misuse is somewhat higher owing to the unique algorithmic power of machine learning algorithms.

We address the issue of misconceptions above in the next Blog of this series where we describe each misconception about AI and how it should be corrected. Regarding the ethical aspect on this item we contend that such ignorant and unfounded fear mongering about AI misconceptions over internet media is highly irresponsible activity intended to generate negative sensationalism only and is therefore a violation of ethics in itself. This is as per the general definition of ethics being a set of moral principles driving responsible activity of any kind.

Regarding the risk of misuse we accept that the dangers of AI/ML being used for criminal activity are real and should not be ignored. Criminal activity howeveris a general risk within human society and there already exists a plethora of regulations and legal measures of multifarious kinds for controlling such harmful activity in the world of civilisation. Therefore in order to mitigate the risk of the occurrence of AI-based crime (examples of which have been deliberately excluded in this Blog) within and outside the corporate world, we recommend and support the following measures to be in place as a matter of high priority

  1. A strong framework of global regulatory prudential standards must be set up for governing the responsible implementation of AI by governments, corporations and individuals. Such standards will support the use of AI/ML for all activities which are directly or indirectly beneficial for human society and these standards will also declare as criminally unlawful the use of AI/ML for any kind of harmful, intrusive, fraudulent, damaging, invasive or other kind of mal-intentioned purpose. Global treaties must be signed by all countries to comply with such international regulatory standards. Google have already proposed a framework for evaluating general purpose AI models against such threats which employ the use of evaluation benchmarks: for more information please refer to the Google Deepmind blog An early warning system for novel AI risks (deepmind.com) .
  2. National Regulatory Authorities need to exist in every country for ensuring the safe and ethical use of AI/ML by organizations. This kind of regulation will include all government departments, large corporations, medium and small companies and not-for-profit In a manner similar to ensuring regulatory compliance within other industries, the National AI Regulator will also require to review regular compliance reports and also conduct audits at given points of time on all organizations implementing AI/ML solutions to ensure that the required standards are being followed.
  3. Within government law enforcement bodies in a country or state such as police or security departments, there needs to be a separate division for detecting AI-based crimes committed by individuals. These divisions need to have in place clear definitions of both the lawful as well as the illegal ways in which an individual can develop and use their own AI/ML capabilities and implementations.
  4. Individual organizations should develop their own AI code of ethicsstipulating clearly to all departments within the organization what kind of AI/ML projects should be worked upon in terms of objectives, tools, technology and data governance. It should also be advised very clearly as to the boundaries (eg, the kinds of data and insights which are being mined and discovered by AI projects) which are in violation of the company’s ethics code and if detected such violation will be regarded as a serious issue subject to disciplinary action. Such an AI code of ethics should be signed by all employees of the organization who are directly or indirectly involved in conducting AI/ML based activity in their areas of responsibility.

 

We now conclude our discussion on ethical aspects of AI stating that while spreading misinformation of a negative flavour about AI is unethical in itself, the risk of AI-based crime occurring does really exist and therefore must be mitigated by having in place compliance standards and regulatory measures at international and national levels, spanning all government departments, organizations and individuals involved in AI/ML of any kind.

Related Posts:

Jan 12,2024

Why is Everyone Talking about AI Today?

While researchers have been working on the various disciplines within the board umbrella of AI for over 7 decades, there existed earlier a major road

By Tirthankar RayChaudhuri

Nov 13,2023

All the False Stories You Have Been Hearing about AI. Science Fiction and Smart Weapons

A number of negative statements which are misleading in nature about the discipline of Artificial Intelligence have been written and circulated in the

By Tirthankar RayChaudhuri

Jan 10,2024

Structuring and Resourcing for a Killer AI Project

In general a Project Manager in Technology is tasked with planning and leading all end-to-end activities of delivering projects for implementing tech

By Tirthankar RayChaudhuri

Jan 16,2024

Linguistics and its Relation to Machine Learning

Text Data - Example The term data (plural of the Latin datum which means a given entity) refers to qualitative parameters and/or quantitative values

By Somsukla Banerjee