|
Carissa Véliz@carissaveliz.bsky.social |
823 followers 140 following 352 posts
Associate Professor of Philosophy at the University of Oxford. Author of #PrivacyIsPower (An Economist Book of the Year). Keynote Speaker. Posts about #privacy, #AIethics, #philosophy, #philtech
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"We’d fallen prey to the comforting illusion that shiny new tech would solve everything. But tech does not work on its own"
"Before we jump headlong into the dream of a humanitarian techno-utopia, we must remember who it is we’re serving and what is at stake." #AIEthics
lithub.com/what-the-ris...
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"If the taking of data without consent is generative AI’s original sin, the widespread labor exploitation is its dirty secret. Lastly, there’s the energy usage, water consumption, and sustainability problems of generative AI." #AIEthics
fortune.com/2024/06/27/g...
0 replies 11 reposts 23 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Leaders need to rethink the traditional “collect-it-all” approach to data." Go for data minimization instead: it'll make your company safer, leaner, and more ethical. #privacy
www.fastcompany.com/91146998/7-w...
0 replies 0 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
It seems that #Apple is doing better than competitors at #privacy with #LLMs, but it's still far from clear whether it's good enough, especially given its partnership with #OpenAI. For the time being, I prefer my phone LLM-free. #AIEthics
www.wired.com/story/apple-...
0 replies 1 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
With retrieval-augmented generation (RAG), a #LLM is hooked up to a database containing domain-specific content, e.g. company files.
"One finding is that better RAG algorithms can improve an LLM more than, for example, making the LLM bigger." #AIEthics
www.zdnet.com/article/amaz...
0 replies 0 reposts 0 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Another tech company completely ignores and tramples over #copyright again. Why? Probably because everyone else is doing it and no consequences have ensued. Is this the world we want to build? #AIEthics
www.wired.com/story/quora-...
0 replies 1 reposts 6 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Too many companies are failing to harness the competitive advantage of #privacy and #AIEthics. "If a company takes clear steps to show they (...) have added in protections, I’m far more likely to try it."
www.pcworld.com/article/2372...
0 replies 0 reposts 0 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Isn't it interesting how two of the most common narratives is that AI is going to take care of the drudgery and leave us the fun parts of work, and that AI is going to take jobs in the creative industry? 🤔 #AIEthics
www.zdnet.com/article/as-a...
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Fascinating. On #TikTok censorship, how it differs from other platforms in its vagueness and arbitrariness, how it infantilizes users and creators, and how it creates new (distorted) language. #AIEthics
thenewinquiry.com/tiktok-llm/
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Meta: it's time to give up on business models based on unnecessary #surveillance. Let it go. Time to move on to more ethical options. #AIEthics #DMA
www.wired.com/story/metas-...
0 replies 0 reposts 0 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Thus far, #LLMs are expensive and not as useful or accurate as some might've imagined. It takes a lot of work to make them work, and they don't always work. #AIEthics
www.wsj.com/articles/ai-...
1 replies 2 reposts 11 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Some police in #Canada are using facial recognition, while others say human rights and #privacy are reasons not to.
The uneven application of the tech—and the loose rules governing its use—has experts calling for national standards. #AIEthics
www.cbc.ca/news/politic...
0 replies 0 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"The lawsuit, filed in a New York federal court, describes OpenAI’s business as “built on the exploitation of copyrighted works” #copyright #AIEthics
apnews.com/article/ai-m...
0 replies 1 reposts 4 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
If you are an American, think about #privacy when you vote.
www.washingtonpost.com/politics/202...
0 replies 0 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Practices like setting prices above the market rates, colluding with rivals, and striking exclusionary deals are unlawful — whether humans or algorithms are behind them."
#AIEthics
www.businessinsider.com/ai-price-fix...
0 replies 1 reposts 4 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Maybe #Meta got misquoted. Maybe they meant to say that AI features that do not comply with #privacy regulations are second rate, and therefore the public shouldn't be exposed to them. #AIEthics
www.businessinsider.com/meta-ai-euro...
0 replies 0 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"At the time Harris was trying to convince #Microsoft product leaders to address the flaw, the federal government was preparing to make a massive investment in cloud computing, and Microsoft wanted the business." 😬 #cybersecurity #AIEthics
www.propublica.org/article/micr...
0 replies 0 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
On the labour behind the AIs, by Madhumita Murgia. "Displaced Syrian doctors train medical software that helps diagnose prostate cancer in Britain. Out-of-work college graduates in recession-hit Venezuela categorize fashion products"
#AIEthics lithub.com/how-vulnerab...
0 replies 2 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Curious solution (to give plaintiffs a share of the company's value), but ultimately not good enough. If #Clearview doesn't have enough cash to pay for the negative externalities it created, it should declare itself bankrupt and disappear. #privacy #AIEthics
apnews.com/article/clea...
0 replies 1 reposts 4 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"P. ¿Firmará The New Yorker un acuerdo de inteligencia artificial con OpenAI? R. No me corresponde a mí decidirlo. Yo creo que ya tomaron toda la información que les dio la gana. Es un escándalo. (...) Actúan con impunidad," dice David Remnick.
elpais.com/babelia/2024...
0 replies 0 reposts 0 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
I don't recommend this long article, but it does contain a kernel of truth: hubris is dangerous, and there's much hubris in the AI world. #AIEthics
"...it is never safe to call on the gods, or even come close to them."
harpers.org/archive/2024...
0 replies 0 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"While #AI often delivers very eye-popping proofs of concept, monetizing them is difficult," said Steve Jones,
www.zdnet.com/article/gene...
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
The US Federal Trade Commission (FTC) has referred a complaint against #TikTok and its Chinese parent company ByteDance over potential violations of children's #privacy to the Department of Justice (DOJ). www.bbc.co.uk/news/article...
0 replies 1 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Generally speaking, we think AI deception arises because a deception-based strategy turned out to be the best way to perform well at the given AI's training task. Deception helps them achieve their goals" And what are those "goals" and who designed them? #AIEthics
www.zdnet.com/article/how-...
0 replies 0 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
I wonder how #Walmart customers feel about their supermarket turning into another personalised ad platform... "Walmart, the world’s top retailer by sales, has discovered that there is serious money to be made in selling access to its customers. #AIEthics
www.economist.com/business/202...
0 replies 1 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
On how a family turned to an encyclopedia to navigate misinformation online.
"Online searches are banished from the dinner table and school projects alike. Instead, my children can go to the encyclopedia for any question they can think of"
#AIEthics
www.publicbooks.org/the-encyclop...
1 replies 1 reposts 4 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
“Part of the value of democracy is to be ruled by representatives who are your peers. And AI is not a peer”. Thanks for the quote, @janestreet #AIEthics www.nbcnews.com/politics/202...
0 replies 0 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Concerns include: cameras are powered by #Amazon (!); slippery slopes; in the document there is no mention of ethics or best practices or minimizing risks or potetnial abuse, and that's how disasters happen. #privacy #AIEthics
www.wired.com/story/amazon...
1 replies 4 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
On the risk of generative #AI rewriting history (not exclusive to the Holocaust, of course). #AIEthics
www.unesco.org/en/articles/...
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Digital sobriety could be part of the actions that people can take as 21st-century consumers and users of this technology." Says @SashaMTL. #AIEthics
www.vox.com/climate/2024...
0 replies 2 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
We are creating systems we do not understand or control well enough, and we are scarily bad at #privacy. Wi-Fi-based Positioning Systems (WPSes) are used by smartphones to learn their position using nearby Wi-Fi access points. They are a privacy threat.
krebsonsecurity.com/2024/05/why-...
0 replies 2 reposts 4 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
¡Gracias!
0 replies 0 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"El simple hecho de ser observados reduce nuestro impulso por experimentar, por preguntar." Ah, y para quien esté interesado, tendremos una segunda edición del curso de ética y IA con la Fundación Rafael del Pino. elpais.com/ideas/2024-0...
0 replies 1 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
“Although they are marketed as something that will enhance your mental health and well-being, they specialize in delivering dependency, loneliness, and toxicity, all while prying as much data as possible from you.” #AIEthics #privacy
gizmodo.com/your-ai-girl...
0 replies 2 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
I agree with the usual worries mentioned here, but also about whether it is ethical to build a machine that is manipulative and deceptive by design because it pretends to be something it's not (i.e., a sentient being) #AIEthics. Article by @kevinroose.bsky.social
www.nytimes.com/2024/05/09/t...
0 replies 2 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
What AI is missing are the embodied emotions that function as our moral compasses, as I argue in my paper on Moral Zombies: link.springer.com/article/10.1...
"People tended to rate AI responses as more virtuous, intelligent, trustworthy, and rational "
#AIEthics
www.bps.org.uk/research-dig...
0 replies 0 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
As predicted, AI is making research worse. The Scopus CiteScore ranking of philosophy journals has 3 journals among the top 10 journals that are fake. They are AI generated, with fake authors, and cite each other extensively to achieve a top ranking.
#AIEthics
retractionwatch.com/2024/06/12/h...
5 replies 141 reposts 190 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
I wonder how many people feel more tempted than ever to get a dumb phone like an old Nokia or a Punkt... #AIEthics
www.nytimes.com/2024/06/10/t...
2 replies 5 reposts 9 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"I didn’t leave the conference feeling confident AI was going to play a meaningful role in advancing any of the UN goals. In fact, the most interesting speeches were about how AI is doing the opposite." #AIEthics www.technologyreview.com/2024/06/04/1...
0 replies 16 reposts 20 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Empowering students with the knowledge and skills to recognise and address ethical concerns about AI is essential. As educators, our goal is to facilitate this process, preparing students to become ethical leaders and responsible innovators" #AIEthics
www.timeshighereducation.com/campus/artif...
0 replies 0 reposts 0 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Each successive generation of the public is accustomed to the privacy status quo of their youth." We need to beware shifting baselines and think more deeply about how much #privacy is needed for healthy democracies, freedom, and wellbeing
#AIEthics
spectrum.ieee.org/online-privacy
0 replies 2 reposts 6 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Many people will think low-cost automated content is a "cornucopian gift," Frank Pasquale said, "until it becomes clear that AI itself is dependent on ongoing input of human-generated works in order to improve and remain relevant"
#AIEthics
www.businessinsider.com/ai-new-dark-...
1 replies 15 reposts 45 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"#Adobe found itself in hot water this weekend after the Ansel Adams estate publicly scolded the company for selling generative AI imitations of the late photographer’s work." #AIEthics www.theverge.com/2024/6/3/241...
0 replies 0 reposts 1 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Because serious confabulation is not bad enough, #Google's AI Overviews will soon have ads. #AIEthics
www.wired.com/story/google...
0 replies 2 reposts 3 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Machine learning models memorize aspects of the training data. That means they are memorizing some personal data. #privacy #AIEthics
theconversation.com/heres-how-ma...
1 replies 2 reposts 6 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"Right now, many companies are wielding generative AI as a hammer, and as a result, everything looks like a nail.Many contexts, such as online support communities, are best left to humans." #AIEthics
theconversation.com/ai-chatbots-...
0 replies 6 reposts 12 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
#Google, further wanting to erode our #privacy.
There are "numerous ways the software could be abused by private #surveillance companies, government agents, stalkers or others who might want to eavesdrop on other people’s phone calls."
www.nbcnews.com/tech/securit...
0 replies 8 reposts 10 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
When asked, “How many rocks should I eat each day,” the tool said, “According to UC Berkeley geologists, people should eat at least one small rock a day,” going on to list the vitamins and digestive benefits.
#AIEthics #LLMs
www.cnbc.com/2024/05/24/g...
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
"The Swedish-based music company Epidemic Sound published a report finding that fewer social-media personalities were using AI tools like chatbots or transcription software in 2024 compared to the previous year."
#AIEthics #privacy #GenAI
www.businessinsider.com/new-data-ai-...
0 replies 1 reposts 2 likes
|
|
Carissa Véliz
@carissaveliz.bsky.social
[ View ] |
Even the name is creepy, #Recall. Do you trust #Microsoft that the data will remain on your computer? Even if the answer is yes (and you're right), what about when you don't have control over your computer (work, airports, hackers, etc)? #privacy
edition.cnn.com/2024/05/22/t...
1 replies 1 reposts 7 likes