
Speaker & consultant helping people implement #ResponsibleAI / #AIethics in their organizations, products & supply chains. https://xitter.gq/t.co/hIaoiyfNIv • https://xitter.gq/t.co/cWqOR16bLA
Seattle
Joined September 2011
- Tweets4.2K
- Following521
- Followers895
- Likes2.4K
It's legal. Most people don't mind. It makes money. Is there a problem? I wouldn't do it on my site—you? cbsnews.com/news/manipulativ… HT @TAPolicy
The author of this piece buries the key focus of AI regulation: accountability for impact of so-called AI technology on consumers/citizens, where AI impact potentially includes not only race discrimination but issues such as privacy & fairness/justice. cnbc.com/2021/03/29/ai-resea…
Once again, using #AI to screen employment candidates sounds great on paper, but real world testing proves otherwise. Don't accept vendor idealism at face value—due diligence is necessary.
Clever assessment of #AI promising to identify personality of job applicants (see below, e.g.: impact of light). Very bad 'publicity' for the provider of the software in question and of an entire industry. Impossible to defend this type of AI upon reading. https://xitter.gq/t.co/v6raJUoKOn
#Privacy. Clubhouse is "starting out as an exclusive network and then gathering the masses into its giant data-hoovering machine" - @jjn1 theguardian.com/commentisfre…
The headline is a little over the top, but well chosen examples within provide nutritious food for thought. venturebeat.com/2021/02/13/t… by @garyg02 #AIEthics
Can brands negotiate app #privacy with consumers? Informed critique of Apple's meta-privacy effort from WaPo's @geoffreyfowler. At issue: brand honesty (or not); Apple's enforcement (or not); differing (& self-interested) definitions of privacy. washingtonpost.com/technolog…
1
New regulation of software "that shape[s] life changing decisions": does a proposed NYC regulation of hiring apps go far enough, or too far?
wired.com/story/new-york-cit… #AIethics
READ this book (by @JanelleCShane). Whether you are an #AI booster or critic, or just a human trying to navigate the 21st Century, it's an entertaining and information-packed look at why AI is so strong & why its limitations can be so counterintuitive: en.wikipedia.org/wiki/You_Lo…
Exciting potential—#AI reading X-rays to end TB. “The machine plus clinician is better than the clinician, and it’s also better than machine alone." - @EricTopol nytimes.com/2020/11/20/healt… by @apoorva_nyc
If you publish an AI "breakthrough" in a scientific journal or conference, do you have a duty to release code that others can run? When is it science vs. advertising (or maybe just telling tales)? technologyreview.com/2020/11… by @strwbilly
2
Makes sense to improve communication with black community + evaluate both demographic inclusiveness of vaccine trials & fairness of distribution. statnews.com/2020/09/21/blac… by @ericboodman #inclusiveness & #technology
Useful perspective for #AI makers / buyers: "what we don't do so much, what our tools don't help us do, is continually ask, who is this [software] for, why are we doing it, and how will people build upon it?" wired.com/story/power-parado… by @ftrain
Bruce Wilson retweeted
Nice long read: #data "should reflect reality, but sometimes reality changes and the data stays the same. That’s why we have notaries, supervisors, lawyers – actually, all those boring people that #blockchain thinks it can do without" thecorrespondent.com/655/blo… #emissions #bitcoin
GPT is "like an improv actor who is totally dedicated to their craft…. Like such an actor, when it doesn’t know something, it will just fake it. You wouldn’t trust an improv actor playing a doctor to give you medical advice." - Douglas Summers-Stay, via @GaryMarcus
GPT-3 == #remix
"What it [GPT-3] does is something like a massive act of cutting and pasting, stitching variations on text that it has seen, rather than digging deeply for the concepts that underlie those texts." @techreview @GaryMarcus
technologyreview.com/2020/08…
Good thread about whether or not AI could be based on an alternative "logic" that is more intuitive to humans than machine logic is.
"using mathematical logic makes our #chatbots less able to have meaningful interactions with humans". This piece thinks Aristotelian logic, based on inferences, is more promising for #chatbots. I know that Aristotle is like a red rag to some, but anyway. theconversation.com/aristotl…
1
"[E]ven without Pestilence, the system is still broken.…" Covid is hastening destruction of already fatally compromised US higher-ed. Key understanding: Universities have been competing over amenities rather than quality of education for decades. Next: start over w/ e-learning.
So everyone is talking about UNC's COVID-19 mess - and all that criticism is perfectly valid.
But we also need to talk about why the uni-administration probably had no choice.
Buckle up and let's talk about university finances and the 4 horsemen of the academipocalypse. 1/lots?
Bruce Wilson retweeted
Historians take note. People are protesting algorithms in the streets.
this is @bencsmoke, Huck’s contributing editor, here in parliament square where students are gathering to demonstrate against the controversial a level results that many received last thursday which saw 40% of students marked down from their predicted grades.
Bravo! "One thing I’ve learned as a scientist is that if you get something wrong, you need to admit it, learn from the experience, and move on. I was wrong." Thank you @StevenSalzberg1 for placing with truth and community over ego—we are in your debt.
I was wrong. Thanks to all those who offered constructive criticism forbes.com/sites/stevensalzb…
Way too much "we better not get caught doing this" going on in this story! Don't rely solely on vendors' claims about #AI. Loop in diversity & inclusion, customer satisfaction, risk management, PR, etc.—not just "loss prevention". reuters.com/investigates/spe…









