More science literacy would be great, although I don't think discussions on social media are the problem. UK energy policy is not being driven by social media, or by anything the electorate asked for; if anything it's being driven by international bodies such as the WEF, the trade union of the very wealthy. Germany's extreme energy policy is an accident of their proportional representation system, in which the Greens (until the latest election) were needed to form ruling coalitions; ordinary Germans quite like nuclear power.
NGOs are highly active on social media. They have some very good ad campaigns, which uses to be physical media based (eg leaflets) but are now mostly online. On the last point regarding German support for nuclear, that's a myth based on questionable polling data. Germans are very anti nuclear. If you ask them whether they think existing power stations should have been shut down, they will mostly say no. But ask them if they want new ones to be built, and they will also say no. Merkel didn't start out as anti nuclear, she caved to public opinion and protests. Take a look at the latest Eurobarometer poll, which gives a more accurate representation of how Germans feel: https://zionlights.substack.com/p/why-doesnt-germany-support-nuclear-the-cloud
I am highly skeptical that AI-based "fact checking" would bring value to important and complex discussions. Perhaps it could address relatively simple falsehoods - like just about any number quoted by Donald Trump when justifying tariff imposition.
But could it provide a clear - and correct - answer to an enthusiast asserting that "solar power is cheaper, and we need more everywhere" Would it point out that the value of solar power depends on many factors - density of the solar resource (i.e. how sunny is it in that place?), grid penetration of intermittent renewables, storage and backup capacity available. And that the price of solar panels is probably one of the least important factors?
I agree there is a distinction. For simple questions it is useful. For more complex reasoning it is still flawed based on the sources it is gathering data from, however I've seen a marked improvement, especially with how it answers questions about nuclear energy. I think it will improve further over time.
the combat and contestation of paradigms is good and appears to occur at all levels, for example with science among practitioners, journalists, and social media. AI is young, but does it seem plausible an AI unit could become relatively non biased? whether combating known biases or collecting known perspectives or the dialogue and criticism among them, AI represents some sort of weighted average of published opinion likely skewed by algorithms, the method of collecting and sorting opinion, bias imported from algorithm authors, and the language pool(s) used for learning. ask any controversial environmental question for example and at least from chatgpt the commonly published degrowth or malthusian falsehood will be the answer. this is emensly time saving to have average views quickly. but it would seem this will always be somewhat behind cutting edge problems, or the bleeding edge at the exploring rim it seems of all knowledge domains where no two experts fully agree on the formulations. the published average would be biased toward more accepted paradigms that are more widely published than emerging contrarian paradigms coming from the accumulation of observed anomalies. the ability for algorithms to acquire combat and contestation might be helpful. is it possible or naive to think that social media is extending competitive thinking to broader groups and these groups are accumulating the ability to discern valid and invalid perspectives, so fact checkers are neither needed nor helpful? so far the western NGO fact checker system on the English based platforms like facebook is 100% biased toward the US/western alliance paradigms, especially seeking to contain facts and observations from the Ukraine and Gaza conflicts. obviously a different system could be made, but how non biased could it be?
More science literacy would be great, although I don't think discussions on social media are the problem. UK energy policy is not being driven by social media, or by anything the electorate asked for; if anything it's being driven by international bodies such as the WEF, the trade union of the very wealthy. Germany's extreme energy policy is an accident of their proportional representation system, in which the Greens (until the latest election) were needed to form ruling coalitions; ordinary Germans quite like nuclear power.
NGOs are highly active on social media. They have some very good ad campaigns, which uses to be physical media based (eg leaflets) but are now mostly online. On the last point regarding German support for nuclear, that's a myth based on questionable polling data. Germans are very anti nuclear. If you ask them whether they think existing power stations should have been shut down, they will mostly say no. But ask them if they want new ones to be built, and they will also say no. Merkel didn't start out as anti nuclear, she caved to public opinion and protests. Take a look at the latest Eurobarometer poll, which gives a more accurate representation of how Germans feel: https://zionlights.substack.com/p/why-doesnt-germany-support-nuclear-the-cloud
I am highly skeptical that AI-based "fact checking" would bring value to important and complex discussions. Perhaps it could address relatively simple falsehoods - like just about any number quoted by Donald Trump when justifying tariff imposition.
But could it provide a clear - and correct - answer to an enthusiast asserting that "solar power is cheaper, and we need more everywhere" Would it point out that the value of solar power depends on many factors - density of the solar resource (i.e. how sunny is it in that place?), grid penetration of intermittent renewables, storage and backup capacity available. And that the price of solar panels is probably one of the least important factors?
I agree there is a distinction. For simple questions it is useful. For more complex reasoning it is still flawed based on the sources it is gathering data from, however I've seen a marked improvement, especially with how it answers questions about nuclear energy. I think it will improve further over time.
the combat and contestation of paradigms is good and appears to occur at all levels, for example with science among practitioners, journalists, and social media. AI is young, but does it seem plausible an AI unit could become relatively non biased? whether combating known biases or collecting known perspectives or the dialogue and criticism among them, AI represents some sort of weighted average of published opinion likely skewed by algorithms, the method of collecting and sorting opinion, bias imported from algorithm authors, and the language pool(s) used for learning. ask any controversial environmental question for example and at least from chatgpt the commonly published degrowth or malthusian falsehood will be the answer. this is emensly time saving to have average views quickly. but it would seem this will always be somewhat behind cutting edge problems, or the bleeding edge at the exploring rim it seems of all knowledge domains where no two experts fully agree on the formulations. the published average would be biased toward more accepted paradigms that are more widely published than emerging contrarian paradigms coming from the accumulation of observed anomalies. the ability for algorithms to acquire combat and contestation might be helpful. is it possible or naive to think that social media is extending competitive thinking to broader groups and these groups are accumulating the ability to discern valid and invalid perspectives, so fact checkers are neither needed nor helpful? so far the western NGO fact checker system on the English based platforms like facebook is 100% biased toward the US/western alliance paradigms, especially seeking to contain facts and observations from the Ukraine and Gaza conflicts. obviously a different system could be made, but how non biased could it be?