The AI 'Black Box' Problem

Fear Not Artificial Intelligence (AI) Bahamas!

By M. Maria Varence
Nassau, N.P., The Bahamas

This is a must needed conversation, but there is another side of that conversation that I pray is addressed.

What happens when we begin relying on systems whose decisions we cannot adequately explain?

This is the AI “black box” problem.

Many advanced AI systems, particularly those built using deep learning, can produce remarkably accurate outputs while making it extremely difficult to determine precisely how the system arrived at a particular conclusion.

Now imagine someone asks a very reasonable question when one of these decisions are made:

Why did the system make that decision about me?

The black box problem exists everywhere, but developing countries face an additional vulnerability.

We are unlikely to build most of the sophisticated AI systems that eventually operate within our economies.

The models may be developed elsewhere, trained primarily on foreign datasets, designed around different populations and regulatory environments and then incorporated into systems used locally.

That creates an important question for countries like The Bahamas:

Are we importing technology faster than we are developing the institutional capacity to govern it?

Consider something as simple as lending.

A foreign-developed AI system may identify correlations that work extremely well within a large North American or European dataset.  But Bahamian employment patterns, household structures, informal economic activity, geography and consumer behaviour do not necessarily mirror those markets.

If we cannot adequately interrogate how the decision was reached, identifying bias or inappropriate assumptions becomes significantly harder.

Much of the AI governance conversation understandably focuses on bias and discrimination.

But black box systems raise another issue: accountability.

If an AI-assisted decision causes harm, where does responsibility sit?

With the international technology provider?  The Bahamian institution that purchased the system?

Or the algorithm that nobody can fully explain?

We cannot regulate an algorithm in the same way that we hold a person or institution accountable.  Ultimately, responsibility has to remain somewhere within the human governance structure surrounding the technology.

That means explainability cannot simply be a desirable feature.

In certain high-impact decisions, it may need to become a governance requirement.

We should not wait until adoption is widespread.

The Bahamas does not need to fear AI.  In fact, I believe small states should be aggressively exploring how AI can help us overcome some of our structural limitations.

But enthusiasm for adoption should be matched by investment in governance.

Before AI becomes deeply embedded we should already be asking:

- What decisions should AI be permitted to make or materially influence?

- When must a human remain accountable for the final decision?

- What level of explanation should a citizen or customer be entitled to receive?

- What capacity do our regulators need to meaningfully challenge these systems?

These questions become even more important across the wider Caribbean, where individual states may not have the technical resources to independently scrutinize AI system entering their markets.

There may therefore be a strong argument for regional cooperation around AI governance.

Developing countries are often encouraged to ensure that we are not “left behind” by technological change.

But there are two ways to be left behind.

One is failing to adopt transformative technology.

The other is adopting technology without developing the institutions capable of governing it.

Source / Comment

Energy is the impediment that will drive the geopolitics behind enterprise tech and AI dominance

There isn’t enough power in the developed world for deployment of these new technologies for even 50% of population, much less AI countries

REMARKS: H.E. Ambassador Professor Gilbert Morris at Bahamas Diplomtic Week Forum on AI and Development

I have only just returned from 10 weeks of lectures and seminars on “AI and the Human Future” in China, Singapore, London, Panama, Mexico and California; which included touring the facilities of the companies and engaging with the individuals who are designing the human future, together with those financing them.

The aim was to both to gain and to explain the actual state of affairs of the coming techno-centric human future:

It’s necessary to state hard facts that are beyond mere opinions or positive or negative perceptions:

1. Only 7% of world’s nations produce any AI tech.  This means 93% have little to no say in AI architecture, design or deployment.  Our definitions of “keeping pace with innovation” usually means someone else’s innovations!

2. In today’s world, that “someone else’s” will be increasingly a non-state actor: Google’s “Gemini” or Anthropic’s “Claude” - Facebook or NVIDA are all larger that 90% of the worlds nations in valuation terms and there is nothing small states can do to get them to alter their technologies for our benefit.

3. Of the 93% of nations left, about 1/3rd are role model states - Denmark, Netherlands, Sweden, Finland, New Zealand etc - with exceptions like India, Singapore and Estonia.  These states don’t manufacture enterprise technology, but they have the means to acquire and manufacture components of enterprise tech, which gives them leverage.

4. The rest are spilt in two; neither of which has leverage: first are mere consumers of tech, whose systems and infrastructure will change to conform to the tech they commune.  Second is the abject “digital divide” where they they will have to make do with dregs and second-hand tech.  They will - by current structures - live off the scraps.

5. ⁠The Promised Balance is Eroding: AI has had diagnostic breakthroughs in health and public-facing innovations in immigration and eGovernment in Singapore and Estonia, for instance.  But the innovations are outpacing the very concept of the nation state, as technology is developing faster than societies, economic and legal systems can adapt.

6. ⁠Social Media now enhanced in impact and speed-of-spread by AI, is producing distortions of reality and neuro-cognitive atrophies; which has reached crisis levels amongst young people and crisis levels in schools and universities.  More than 2 dozen nations have banned social media access for children at school or between 7pm and 7am; with China limiting social media to 2 hours per day for children.

7. ⁠At the heart of social tech is personal identity: Estonia solved this by declaring that a citizen is now the unity of one’s Biological and Virtual self; a unified identity.  Estonia also made internet access a Human Right in their Constitution; additionally, Estonia held their last 8 General Elections online for 1.3 million people.  In Estonia, AI does not replace people but is structured and used to enhance human life and social systems.

URGENT PRIORITIES:

1. We need humane technologies that prioritises humans and enhances human living systems rather than render humans useless.  Some sociologists argue that we are too divided for to generate the needed solutions.  But I find that people are not divided: “they are being divided” for the benefit of a few tech companies.

2. ⁠Social mapping, Economic Mapping, Health and Payments tech, Plurality Systems aimed at enhancing political participation are the best uses for AI and enterprise technologies.

3. ⁠The problem - AGAIN - is, small nations in the 93% don’t create this tech and it’s being designed by those creating it for human control rather than human enhancement.

4. ⁠Small nations must urgently align and cooperate to ensure tech deployment at a pace of social absorption, without the ill-effects of limbic-colonialism or dopamine lust!

5. ⁠Small nations must convene a global congress aimed at setting new rules that will change our economic models across the world more in the next 3 years than the last 1,300 years.

6. ⁠But there are additionally two main crucial matters:

a. Even if managed well, AI will induce massive labour destabilisation.  It’s not enough to say it will create jobs because AI companies will create AI to take those jobs too.  Some advise a switch to Universal Basic Income (IBI).  But UBI can’t be universal: manufacturing nations would have no UBI whilst non-manufacturing nations would have it, so as to create a market for transactions.  This needs to be mapped out immediately.

b. There isn’t enough power in the developed world for deployment of these new technologies for even 50% of population, much less al countries.  Energy is the impediment that will drive the geopolitics behind enterprise tech and AI dominance. 

Source / Comment