[DC] Quality Plan for AI is a Must

[Amali De Silva - Mitchell]

Dear DCs

We are reading so many stories about AI and what will happen to the world
and so forth +/ -. Some have asked to stall development. For healthcare it
AI needs to keep developing to reach UN SDG 3.

This DC has maintained that AI can be very beneficial, if it is Quality AI
but detrimental if poorly sourced or progmmed AI.

This is to the extent that for medical aps a "random" error caused from
trojan data can impact a life to death.

There is a need for the UN to bring together a "Quality Plan for AI'. This
plan will mandate the quality of data going in, the processes, algorithms,
including mandated accessibility, safe guards on outputs, continuous
monitoring and so forth.

As a CPA member,  our body has what are called peer reviews of audits and
audit company management, so as to maintain a high level of professional
quality and ethics. This is the kind of concept in mind here. However, due
to intellectual property issues an independent body such as the UN can
carry out these activities. An example at the UN is the Nuclear audits body
etc.

AI developers must also be mandated to take ethics courses and understand
what are not the common cultural values.

Data shared down stream from other feeder aps must also comply. Algorithms
must be mandated by law to be audited by not just one group, but multiple
groups depending on the nature of the ap. Some AI aps could take a simple
audit.

The UN can be instrumental in managing 'Quality AI as a Human Right and of
Industrial Peace Keeping'.

Perhaps there is a need to partner with the WTO and the other scientific UN
agencies including climate change for zero climate impact AI as our DC has
advocated.

The Speed of Trust (Steven Covey; DC session at IGF 2020) should be the
rate determining step for the development of AI.

Trusted processes and frameworks that are more detailed must be developed
and made into a UN  treaty. ISO and other bodies cannot enforce, but a
treaty can.

There must be an ombudman set up etc. The world is made up of indiciduals
and thr individual voice must also be heard. This can be a structures body
similar concept to NRIs and IGF.

The UN is in the place to make this happen!

If the quality of AI is managed, the speed of trust will follow, which must
also have excellence in muticultural based ethics (each culture has its own
renditions) and a common base of values for actual application development
and operation. Includes privacy etc.

I plead with this body to promote Quality AI  with the Internet.

Quality Data is Critical (relevant, timely, non-skewed, representative,
evidenced etc. )

The 'Speed of Trust' process will give us better AI outcomes at the right
pace, the Quality, Ethical pace.

Kind regards, thank you for listening.
To great outcomes!

Amali De Silva - Mitchell
Coord IGF DC Data Driven Health Technologies.

Amali De Silva - Mitchell MSC., CPA, CMA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://intgovforum.org/pipermail/dc_intgovforum.org/attachments/20230610/4aed97e7/attachment.htm>