This is interesting because it has the potential to address systematic barriers to entry and also does not require candidates to give up on stable income to get into teaching. From the article, I see that it is the very conditions that create discrimination in the classroom, that continue to be reproduced because they have not been addressed at the root-cause.

I quite honestly find it ironic that the writer is placing the U.S. as an example here. It would make more sense to one compare to a country that does not have its education system crumbling, and also one with a more similar governance context.

Building on the earlier point of how Canada's governance systems influences how strategies, policies and guidances are put into place. Because of the provincial nature of government, when a democratic event happens and there is a change, it can often mean overhauling bad policies but also good ones too. At the federal level Canada has "Health Canada", should it not consider similarly "Education Canada" and should there not be checks and balances in place to maintain equitable and just access to an education for all?

This is a powerful statement. If diversity is commodified and KPIed, how do we know what it truly means and what impact it has on student success? Do we have enough evidence for current DEI efforts to know whats working and whats not. More importantly, to what extent do students and parents have their voices heard by the board to truly measure if diversification is positively impacting the communities. This is perhaps a case example of where quantitative indicators cannot possible tell the whole story and a qualitative study would definitely be required.

While teacher diversity is part of the solution, its one that takes time to enact especially when there is a national level teacher shortage. What about teachers already in the classrooms? What is being done to resolve some of the critical issues they may be facing, and even creating, in classrooms?

I do not understand why, but Canada seems to shy-away from strategy in many regards. There seems to be less appetite for big picture commitments, and more willingness for light touch "guidance". I think this can be largely owed to Canada being a parliamentarian system, which means there can be often be many changes in government. Hence, long-term strategies may flip-flop but guidances have a stronger chance of survival against political shifts.

True - while Canada can be considered a racially stratified society, I would argue many other countries can be more strongly considered to be "stratified". Other countries, especially those who have economic/labour systems that are literally based on the nationality of citizens. While Canada remains socially stratified, priveleging "whiteness", in many respects it has made strides towards equity far beyond many other countries.

This is why the work of Gebru and the likes is so important, they are the organic intellectuals who have nothing to gain from speaking up, but much to lose. That's why we must learn even more so to these voices, because surely there is something within them that is pushing them to speak up so strongly against this.

This is exactly it - AGIs are designed to re-create and enforce hegemonies of the global elite and those in power. Its designed to continue to uphold existing systems and power dynamics; and it will require the re-imagination of how AGIs can positively contribute to societies and invent competing/parallel systems that have a stronger motivation than to be a money-machine to keep them alive to serve the public good.

It really boils down to this. What and who can we trust? While yes its often referenced to the world order and global political stage - it often applies at the individual and community level when people do not know what to believe or who to believe. Mis and dis-information has spiraled even further when there is a tool i.e. AI that can make any lie appear to be true. And unless their is a society that has an understanding on how to investigate that it will not know how to navigate. Although, knowing truth from lie will not in any way resolve trust, it might make it even worse.

Perhaps this is a useful use of AI - using it as its own correction model. If AI can be used to upload and enrich online sources of various languages and of diverse sources - and then be trained on those sources - then that might be a way to get AI to be more equitable and just in the output it creates.

So even in the rare cases where AI says something sensible, merely because the LLM it was trained on had enough evidence in that direction - we see that it can be completely over-ridden by those who own it. It makes me wonder if there is a need for non-profit driven AI startups, the likes of wikipedia that are just a public good. But unlike wikipedia which just requires a laptop and a person to contribute with their intelligence; AI needs to be powered... so unless by some cause of philanthropy or some revolution that allows AI to function without consuming so much energy, I doubt this will happen with ease.

I find the term "stochastic parrots" truly interesting. Although its used in reference LLMs, when I read Gebru's paper I found my mind wandering off to many things. But mostly just how much we have seen this patten before where a brave person speaks against the hegemonies created by the rich and powerful. Where they are pushed out and things continue as is - if not worse. And so I wonder how many of us choose to be "stochastic parrots" on the day-to-day because it keep those around us happy and us out of trouble.

Building on my earlier points, we see this re-enforcement and regurgitation of world views, knowledge, values of the ruling class re-enforcing their positions of power. We see that even when every day people have managed to find somewhat of a balance with their AI usage/exposure, it is now being enforced by institutions (government, workplaces etc.). This questions how much agency citizens truly have in this AI world? How much control do we have - not only on our personal data but also in what we are exposed to in this constant influx of information and suggestions by AI? What is our ability to shift this i.e. make revolution with/in AI contingent upon?

This in my view describes GenAI very accurately. We are seeing cultural hegemony - previously determined by SEOs etc. - now being replaced by AI. However, this is a case where rather than a technology serving as a "great equalizer" or perhaps "brining the world closer together" its one where society is going down a slippery slope of deep fakes, mis/dis-information, plagiarism amongst the many plagues. Beyond that, society is still grappling with solutions to this. And with a technology that is moving at an extremely fast paced, it is not until there is a slow-down that society might have a chance of catching-up with potential antidotes.

While the idea of alternative LLMs is hopeful, it also brings to the surface some of the key issues that also arise when looking at knowledge on the internet being predominantly in English. LLMs can also be trained on existing knowledge, and hence there is a double tax to pay here. Whether its knowledge from other sources (ex. Global South) or in another languages (ex. Mandarin, Arabic Etc.) this means this knowledge needs to exist in the first instance for LLMs to be trained on it. And hence there will be an advantage to the model that is trained with the most amount of resources - resources that themselves possess a bias.

How does one engage in this "war of position" when the algorithm now has power over transforming culture, meanwhile it is the instiutions behind the algorithm that determine what is worthy/notable/important. As we look at the approach of Gramsci in the AI-era, does one now engage in war of maneuver and that of position by feeding the AI machine and figuring our the every-chaging algorithm. Or is there another way of engagement.