#RiseOfAI@Riseof_AI starts with a chatGPT piped to text-to-video keynote. @bootstrappingme says 99% of people can't spot that this is not a person. Are there stats about this? I had about 0.2 seconds of doubt, but I'm sure there's better text to speech out there now.
Actual costs of regulating high risk AI under the #AIAct is negligible, see: ideas.repec.org/p/osf/socarx/8…
People from any sector but digital is laughing at the whining about these costs, many will happily tell you that compliance with regulation often IMPROVES your product. #riseofai
Whoa, weird claim #riseOfAI that we all want "the same education, the best education". 1) what's best varies by student (capacities, interests, where they can live) 2) the last thing humans need to be in the age of AI is identical to each other! 3) AI can help customise!
1/2
I'm not too worried about universities going away, we're still useful, though on the other hand I do worry for ALL jobs that people use the period of change to undermine employees hard-won rights (and wages) joanna-bryson.blogspot.com/2016/01/what-a…#riseofai#AIEthics#FutureOfWork 2/2 on edu
That was @hansuz#riseofAI he's giving us a really interesting account of Chinese AI regulation right now. But he expressed surprise about people expecting more accuracy from AI than its data. Such mistaken faith in machines as more reliable than humans is called #automationBias.
Hans–GPT can't have AGI because of lack of multi-step reasoning. IMO to some extent culture stores & we mine those though, that might what's surprising some AI researchers. But agree that without ape motivations/phenomenology you don't get humans wired.com/story/lamda-se…#riseofai
I've been super interested in quantum for NP-hard problems for decades. A lot of symbolic/explainable AI could become tractable if that ever actually worked. But last I knew quantum energy costs were so extreme it was unlikely to be used except for security. #riseofAI@QuantumViv
Oh, Sabina Jeschke just brought up the canonical NP-hard problem (traveling salesperson problem) as an application of quantum for logistics, apparently they are running some sims of this now. #riseofAI
Jeschke seems to be trying to undermine arguments that quantum is intractable due to sustainability issues by claiming quantum-fueled AI will solve sustainability. I hope she's right, with or without the quantum part, we're starting WAY to late on social adjustments #riseofAI 1/2
I'd love to hear a #riseOfAI talk about using AI to improve human governance and persuasion to such an extent we can get people to adequately stop flying, using cars, doing other ecologically destructive things.
I'm really getting angry about of irresponsible anti-#AIAct misinformation @Riseof_AI#riseofai. E,g. "we're about to see the first version of the AIA" no, it's about the 5th, and FINAL. "they didn't think about startups" NO that was the key concern.
I can assure you, every one of the many consultation sessions I've been in on the #AIAct talked about both what worked & what didn't wrt the #GDPR, and especially the need for #enforcement (info for @ZennerBXL#riseofai) So I agree in part.
@ZennerBXL I'm actually OK with politicians taking credit for other people's work, it's apparently an important part of democracy and legitimacy. I only get angry when they undermine that work with false critiques. Oh, q&a, now what will I say, or do I wait for my own keynote? #riseofAI
Interesting Q: will people really want to buy EU AI (weirdly at the end the panel u-turned to saying people will want to buy ours after the AIA) if their values are different. "value aligned" really should mean "aligned with owner/operator" IMO #riseofai
OK I ranted. @ZennerBXL describes another late hack I strongly agree is bad: shoehorning extra human rights & sustainability baggage only on AI. It should be done like liability, updates for HR & sustainability should be general to all products, & carefully legislated. #riseofai
Sorry I'm not tweeting as much, but great talk by @Deloitte's Björn Bringmann AI IN EUROPE – THE GOOD NEWS. Germany is actually the 4th largest national AI economy (!). Similarly Thomas Neubert on the present panel is talking about how startup culture is booming in DE #riseofai
Neubert: "I only wanted to develop until 5 months ago, but the EU opened an office in SF now and I go there, and now I get it. Only Europe is fining Google & Microsoft when they wrong.I do believe the EU has the right and are the best positioned to regulate. Even US sees that."
chair (otterbach, great name!) what if youth are afraid of all these regs & don't make startups? Tina Klüwer we need to remind people of all the awesome stuff we already have that is AI. We talk too much risk and not about all the wins. LIke all the stuff on your phone. #riseofAI
"I'm glad there's no Elon Musk in Germany, but moving more towards the area of AI" ha ha ha ouch. #riseofai
@hansuz actually in China they hate cameras too and often obscure or interfere with them. #riseofai
Q why people fear science, don't believe in progress. My own answer would be this is a form of political polarisation, which from the research I've been doing looks to be correlated with economic precarity–justified fear. cf my pinned tweet, or joanna-bryson.blogspot.com/2022/02/scienc…#riseofai
• • •
Missing some Tweet in this thread? You can try to
force a refresh
"The problem is that we in the tech industry don’t like regulation. Not because we’re evil but because we know that regulation tends to make a mess of things. We like the threat of regulation and we hope that it will keep things at bay without actually requiring stupidity." 1/2
That's @zephoriazephoria.org/thoughts/archi… I assume it's why tech is panicking, because they are used to only being threatened with regulation, and they think it can't be done well, despite having water, sewage, electricity, &c. in their homes.
Moved fast, broke things, cops came.
@zephoria There is, incidentally, shockingly little about utility regulation on the Internet. It's all old. It's like it's just for stuff we recognised in the early 20C or something. There couldn't be new utilities.
Except software seems to be making every industry into natural monopoly.
Then that same #inequality drives #polarization by creating false scarcities, which produce economic threats, which make lower-risk (but lower-payoff) opportunities more attractive than others. This part of the equation is what our new paper is about.
Just found my last talk before self isolating then lockdown last March, which went well! NATO @STRATCOMCOE asked me to talk about social media & #democracy which isn‘t my area, so I got to talk about what were IMO the most important results of others.
So this is happening. I have no idea what I'm getting myself into, other than some idea of who I'm getting into it with. I'll certainly do my best to do good. blog.google/technology/ai/… #aiethics
If you want to know more about my take on ethics, you can read my Ethics of AI web page: cs.bath.ac.uk/~jjb/web/ai.ht…
Or just watch this film, which pretty much covers the gamut of AI ethics, despite being in the context of cars. cnnmoney.ch/shows/front-se…
Here is what I currently think is the biggest IMMEDIATE issue in technology policy feed.itsrio.org/computational-…
And here is a VERY recent AI policy book chapter by me that talks a bit about tech giants, as well as the nature of AI bbvaopenmind.com/en/articles/th…