From the Opening of the Global Institute of Strategic Research (GISR), HBKU, Doha
Ikhlaq Sidhu, Dean, School of Science & Technology, IE Madrid
It was pleasure to participate in the opening of the new Global Institute of Strategic Research (GISR) on October 1, 2023.? This Institute's content was brilliant and high quality.? Congratulations to the GISR leadership, HBKU, and the Qatar Foundation for this amazing contribution to the world.
I attended the AI & Governance Roundtable.? The panel included Michael Wooldridge, Oxford professor and author, Carnegie Mellon President Farnam Jahanian and with the Qatar's Minister of Communications and Information Technology, Mohammed Bin Ali Al Mannair, and Dr. Sanjay Chawla, Director of the Qatar Center for Artificial Intelligence.
Note, at the same event, the other 2 roundtables were Future of Energy and Implementing Sustainable Agreements with Fragile States. ? ?
These are my insights and observations oriented about what is new to the AI & Governance conversation:
Lets start with the baseline conversation of the accelerators vs detractors of AI:
The Accelerators of AI are:
- AI is on a very accessible platform, ie the world wide web
- It has a natural interface, and
- AI does not need a new business model to use it, its incremental to everything we already do
The Detractors of AI from a policy view are:
- Results are not explainable
- Illicit usage including misinformation
- A lot of power in held by a few private companies
Given this starting point, here are 7 interesting AI policy points:
- No one said this directly, but there is an inherent trade-off between how much we regulate (limit usage in a local region) versus the global competitiveness (of that local region).
- Jobs: Yes,? knowledge-worker jobs can be disrupted, but I felt that this part of the discussion was “old world thinking” or at least imbalanced.? This is because of the new larger scope of work that will be enabled (like designing a full personalized product unique to a single user, instead designing only one feature) So far, the cost of these new scopes of work have been too expensive.
- Culture: On a global scale, many countries are concerned that the language of training of LLMs has so far been English, and that as a result, cultural norms are being set to biases from that subset of English speaking, western cultures.? To me the question is whether this technology is pushing us all towards one “averaged culture” or will AI be the translator across cultures that we can all speak with.
- AI Decisions and Regulation: A jobs example came up, which was that call center workers mostly read scripted responses, so they are likely to be replaced by AI.? And at the same time, another conversation stated that we need to keep a “person in the loop” so that AI does not make decisions without being checked by a human.? The interesting part of this is that there is not much value an AI which cannot make a decision on its own.? So again, this is only my perspective: but one of the easier knobs to regulate AI is actually the enablement of autonomous decisions.? If AIs are not allowed to make decisions, they are effectively not very valuable.? The field and scope of the restricted decision set is a powerful regulatory tool in my opinion.
- Misinformation: One of the larger concerns was that AI will become a mass misinformation tool, and in a manner we have not seen so far.? Partly this is because we cannot distinguish AIs from humans in conversation.? AI can seed fake news at scale, and a very low cost structure.? If a typical misinformation campaign costs approximately? $10M USD or Euros, the new AI based campaign may cost only 10,000.? This is empowering to new small bad actors.? A follow-on question is whether AI generated content can be forced to be “marked” or “identifiable” as AI.
- Cross-Border AI: On the global regulatory AI perspectives: ?USA policy is based on stopping it when harm is detected. ?The European policy is follows a rights based approach. And, China is using a government assistance focused approach. We understand that there has been increased separation between US and Chinese interests, but the new conversation was that there are two opposite forces at play: a) the US would like to limit China’s technology aspirations, at least those connected with US companies vs b) the US is also interested that US companies have access to Chinese markets. Therefore, in the end, there will likely be some equilibrium.
- Policy Regulation vs Technology: A big take away for me was that regulation should not be technology specific. That can be over-prescriptive and too narrow.? Technology moves so quickly that regulators can get caught writing and rewriting regulations and basically have a regulatory mess that no one can understand or implement. For example, GDPR was written prior to ChatGPT. ? It is more effective to regulate for the effects and safeguards for its people.? This is a lesson I’d like to take with me to the our newest project, the IE Ripple Virtual Asset Regulation Lab.? Is it possible that the regulation of AI and the regulation of Virtual assets might have a common, non-tech ,policy core. ?
Appreciating the Global Institute of Strategic Research (GISR). HBKU, Doha.
Some policy makers focus on the global voices and issues. Other policy makers focus on the domestic voices and issues. When these two groups don't have common information, it results in mixed messages and bad policies. GISR was developed to be a bridge!?
GISR is an independent and interdisciplinary think tank concerned with regional and global challenges of government, advancement, and peace.? GISR provides the research and insight needed to support smart decision-oriented policy makers in the Middle East region. ?Any country or region would benefit to access to an Institute like this.
Overall, this event was an amazing opportunity to learn.? The quality has been excellent. Congratulations once again to HBKU, and the new Institute, HBKU, and Qatar.?
PS: A shoutout to Blue Hydrogen and Medibuddy, per our side conversations.
