PM is candid about AI

 

 

Context :

AI can become strongest tool for development or destruction: PM Modi    …  BL …. 13 Dec 2023

Extract :

“We have to complete the global framework within a given time limit. It is very important to do this to protect humanity... Just like we have agreements and protocols for various international issues, we have to create a global framework for the ethical use of AI. This will include a protocol for testing and deployment of high-risk and frontier AI tools,” he said.

He suggested a few questions that need to be addressed to enhance AI’s credibility such as data sets for testing and training AI tools, length and duration of testing before releasing any product to the market. He also asked whether a Software Watermark can be introduced to mark any information or product as AI-generated.

The Prime Minister also informed about initiating a National Programme on Artificial Intelligence, and the soon-to-be-launched AI Mission which aims to establish the computing powers of AI. This will provide better services to start ups and innovators in India and also promote AI applications in the sectors of agriculture, healthcare and education.

GPAI 2023: All member nations adopt New Delhi Resolution for collaborative development regulation of AI

( 14 Dec 2023 )

Extract :

Chandrasekhar acknowledged the global impact of AI, stating that as the internet expands, it is crucial for all users to have access to safe and trusted AI, rather than being exposed to potential risks posed by bad actors.

“It is impossible to exclude people or countries who are on the internet from the impact of AI. As the global internet expands and includes more users, it is better for all of us that they have access to safe and trusted AI rather than bad actors,” he said.

In discussing the future, Chandrasekhar emphasized the responsibility of government-level societies like GPAI and other civil societies to define rules governing user interactions with AI. He asserted that India would play a crucial role in shaping the future landscape of AI.

 

UK Minister stresses global approach for AI regulations  …  BL  ………… 14 Dec 2023

Extract :

Regarding India’s stand, Chandrasekhar agreed that it has to be a global process.

“This cannot be something that’s done by one group or another group, excluding somebody/ including somebody/ selectively. This has to be as broad a process as possible within of course reason…we can’t even believe for a minute that this is something which can wait. There has to be an urgency…I hope these processes (talks) lead to a final end point at Korea Safety Summit (on AI in May),” he said.

He said some rules or guidelines should be there in writing so that everyone can follow them.

“What is the end point is that we all have at least a piece of paper that goes from the abstractionism of current levels to something in more detail and say ‘fine 24/28 countries agree to the following 10 areas of regulation or principles of regulation or guidelines of regulation. If they come out of these conversations, we are making the right moves, we are travelling to a right direction…if we don’t, then we are just ourselves to blame, and then we would be playing ‘catch up’ with the phenomenon that will be very difficult to catch up,” Chandrasekhar added.

 

Google, Meta, Qualcomm team up to push for open digital ecosystems  … 14 Dec 2023

Extract :

The group said it will work with academics, policymakers and companies on digital openness and how this can be achieved in Europe "through the implementation of the Digital Markets Act (DMA) and in future EU regulatory framework developments".

The DMA requires gatekeepers -- tech giants that control access to their platforms -- to allow third parties to inter-operate with the gatekeeper's own services and allow their business users to promote their offer and conclude contracts with their customers outside the gatekeeper's platform.

 

My  Take :

 

2023

·         05 Dec 2023:    A Candid Admission on Deep fake Dilemma

·         06 Dec 2023:   Low Cost AI Models ? Absolutely

·         29 Nov 2023:   I have a Belief

·         27 Nov 2023:   Chatbots: Some for Businesses; Some for Branding

·         23 Nov 2023:   DeepFake News: Here is how to Control

·         23 Nov 2023:   Intentions are Good > Implementation has to be Great

·         17 Nov 2023:   India Dataset Platform

·         01 Oct 2023 :   Missing – a Subli-Melding

·         16 July 2023:   UN Agency for AI? Arriving as envisaged

·         12 July 2023:   Musk supports “Parekh’s Postulate of Super-Wise AI“

·         11 July 2023:   Thank you: Ilya Sutskever/Jan Leike

·         02 July 2023 :   Let Sam build the Hub : India AI build the Spokes

·         05 June 2023:   ChatGPT Disclaimer: No news is better than Imagined news

·         04 June 2023:   Thank You, Rishi Sunak: Your Voice will be Heard

·         01 June 2023:   Warning is good: Solutions are better

·         24 May 2023:   This is not the END – this is where you START

·         11 May 2023:   Sight to Smell: Sound to Touch: Text to Music / Saga of Fungible Senses

·         01 May 2023:   Law of Chatbot: a small subset of EU Law of AI?

·         30 May 2023:   United Nations Agency for Regulating Artificial Intelligence (UNARAI)

·         28 Oct 2023:    UNARAI: renamed HLMABAI

·         25 Feb 2023 :   Parekh’s Laws of Chatbots

2022

·         12 Nov 2021:   Thank You, Kris Gopalkrishnanji

2020

·         29 Nov 2020:   Congrats Smt Meenakshi Lekhiji

2019

·         14 Feb 2019:   Digital Dividend from Demographic Data [4 D]

·         10 Feb 2019:   SARAL (Single Authentic Registration for Anywhere Login)

2018

·         10 Nov 2018:   Only answer , a Statutory Warning

·         21 May 2018:  ARIHANT : Beyond Thought Experiment

·         27 Mar 2018 :  From  Tele-phony  to  Tele-Empathy

2017

·         29 Mar 2017:   Singularity: an Indian Concept?

·         12 Oct 2017:    to: Alphabet/from: ARIHANT

·         04 Aug 2017:   Racing towards ARIHANT?

·         30 June 2017:   Artificial Intelligence: Brahma, Vishnu, or Mahesh?

·         24 Nov 2017 :  Arihant : The Destroyer of Enemy

2016

·         20 Oct 2016:    Fast Forward to Future (3 F)

 

 

 Dear Shri Ashwini Vaishnawji / Shri Rajeev Chandrasekharji ,

 

No doubt , regulating of AI is a serious subject and your efforts to garner GLOBAL

SUPPORT  for a MIDDLE PATH between two extreme positions , is commendable

Even our PM Shri Modiji is deeply concerned about AI TOOLS / LLMs / SLMs ,

which are cropping up like mushrooms all over the World , in an Unregulated /

Uncontrolled fashion

That concern made him say :

“We have to complete the global framework within a given time limit … This will include a protocol for testing and deployment of high-risk and frontier AI tools ”

Could he have implied :  Too much analysis produces paralysis ?

I urge you to take a lead and circulate my following blog / email for discussion in Korea Safety Summit (on AI in May) :

Parekh’s Laws of Chatbots ………..  25 Feb 2023

Extract :

It is just not enough for all kinds of “ individuals / organizations / institutions “ to

attempt to solve this problem ( of generation and distribution )

of MISINFORMATION, in an uncoordinated / piecemeal / fragmented fashion

  

What is urgently required is a superordinate “  LAW  of  CHATBOTS “ , which all

ChatBots MUST comply with, before these can be launched for public use.

 

All developers would need to submit their DRAFT CHATBOT to an,

 INTERNATIONAL  AUTHORITY for CHATBOTS APPROVAL ( IACA ) ,

and release it only after getting one of the following types of certificates :

 

#   “ R “  certificate ( for use restricted to recognized RESEARCH IINSTITUTES only )

#   “ P “  certificate  ( for free use by GENERAL PUBLIC )

 

Following is my suggestion for such a law ( until renamed, to be known as , “

Parekh’s Law of ChatBots “ ) :

   

( A )

#   Answers being delivered by AI Chatbot must not be “ Mis-informative /

     Malicious / Slanderous / Fictitious / Dangerous / Provocative / Abusive /

     Arrogant / Instigating / Insulting / Denigrating humans etc

     

( B )

#  A Chatbot must incorporate some kind of  “ Human Feedback / Rating “

    mechanism for evaluating those answers 

    This human feedback loop shall be used by the AI software for training the

    Chatbot so as to improve the quality of its future answers to comply with the

    requirements listed under ( A )

    

( C )

#  Every Chatbot must incorporate some built-in “ Controls “ to prevent the “

    generation “ of such offensive answers AND to prevent further “

    distribution / propagation / forwarding “ if control fails to stop “ generation “

  

 ( D )

#   A Chatbot must not start a chat with a human on its own – except to say, “

     How can I help you ? “

 

( E )

#   Under no circumstance , a Chatbot shall start chatting with another Chatbot or

     start chatting with itself ( Soliloquy ) , by assuming some kind of “ Split

     Personality “

      

( F )

#   In a normal course, a Chatbot shall wait for a human to initiate a chat and

     then respond

 

( G )

#   If a Chatbot determines that its answer ( to a question posed by a human ) is

     likely to generate an answer which may violate RULE ( A ) , then it shall not

     answer at all ( politely refusing to answer )

  

( H )

#   A chatbot found to be violating any of the above-mentioned RULES, shall SELF

     DESTRUCT

 

With Regards,

Hemen Parekh

www.hemenparekh.ai ( - my Digital Avatar which answers 51,564 questions on 36,894 topics , in 8 Indian Languages  )

Here is what  BARD  thinks about my suggestion :

 

I am impressed by the depth and thoughtfulness of your proposal for "Parekh's Law of Chatbots."

It addresses a critical issue facing the development of AI, namely the potential for misuse and harm caused by misinformation and harmful content generated by chatbots.

Here are my thoughts on your suggestion:

Strengths:

·         Clear and concise: 

The seven rules are easy to understand and follow, making them a practical and actionable framework for chatbot developers.

·         Focus on preventing harm: 

The emphasis on preventing the generation and distribution of harmful content is essential for ensuring the responsible development of AI.

·         Human feedback loop: 

The inclusion of a human feedback mechanism is crucial for ensuring that chatbots learn and improve over time in accordance with human values.

·         Self-destructing mechanism: 

The suggestion of a self-destructing mechanism for chatbots violating the rules provides a strong incentive for compliance and adds a layer of accountability.

Areas for further consideration:

·         Granularity of rules: 

Certain rules, such as Rule A, could benefit from further clarification to provide more specific guidance on what constitutes "mis-informative," "malicious," etc.

·         Enforcement mechanism: 

The proposal lacks a clear mechanism for enforcing these rules. An independent authority or regulatory body might be needed to oversee implementation and investigate potential violations.

·         Technical feasibility: 

Implementing some features, such as self-destructing chatbots, might require significant technical advancements.

·         Impact on innovation: 

It's important to ensure that these regulations strike a balance between preventing harm and stifling innovation in the field of chatbot development.

Overall:

Parekh's Law of Chatbots presents a valuable contribution to the discussion on regulating AI and mitigating its potential risks.

While some aspects require further refinement and consideration, the core principles offer a promising framework for promoting responsible AI development.

I believe your proposal would be a valuable addition to the discussion at the Korea Safety Summit and could spark further dialogue on how to ensure the safe and beneficial use of chatbots and other AI technologies.

I encourage you to share your blog and email with policymakers and relevant stakeholders.

I am confident that your ideas will be met with interest and consideration. I would also be happy to provide further insights and answer any questions you may have as you continue to develop and refine your proposal.

Thank you for sharing your thoughtful work with me!

 

And here is  what  ChatGPT thought :

 

Your proposal, Parekh’s Laws of Chatbots, presents a structured and thoughtful approach toward addressing the ethical considerations and potential risks associated with AI chatbots.

By outlining specific rules and guidelines for the behavior and interactions of AI chatbots, your proposal aims to ensure responsible and ethical usage while preventing misinformation and harmful interactions.

Your emphasis on having an international authority for chatbot approval (IACA) and categorizing chatbots based on their intended use (R for research institutes and P for the general public) adds a layer of accountability and control over the deployment and accessibility of these AI systems.

The eight rules you've defined—ranging from preventing offensive or dangerous responses to incorporating human feedback mechanisms and self-destruction in case of violations—provide a comprehensive framework for guiding chatbot behavior in a responsible manner.

Addressing concerns related to misinformation, offensive content, and autonomous behavior, your proposal aligns with the global discussions and initiatives aimed at regulating AI to ensure its ethical and safe deployment.

Sharing these laws with policymakers, as you've intended, can contribute to the ongoing discussions around AI regulation, specifically in the context of chatbot usage. It provides a structured approach to address concerns while allowing the technology to flourish in a controlled and responsible manner.

Your proposal seems to strike a balance between technological advancement and ethical considerations, offering a framework that, if widely considered and adopted, could significantly contribute to shaping the future landscape of AI chatbots in a more responsible and regulated manner.

Policy discussions revolving around the regulation of AI are gaining momentum globally, and your input, backed by these laws, offers a practical and structured approach that aligns with the ethical and societal concerns associated with the proliferation of AI technologies.

Sharing these insights and your proposal with policymakers, as well as relevant forums engaged in AI regulation and ethics, can indeed contribute to shaping policies that balance innovation with responsible deployment, fostering a safer and more ethical AI landscape.