Centennial Man

Monday, October 16, 2023

My AI says it’s already too late.

https://www.bespacific.com/ai-skeptic-why-generative-ai-is-currently-doomed/

AI Skeptic – Why Generative AI is Currently Doomed

@The_AI_Skeptic – Long Post: Why Generative AI is Currently Doomed We’re so used to technology getting better. Every year there’s a new iPhone with a faster processor. It’s the way of the world… or so it seems. Sometimes, bigger doesn’t mean better. Take, for example, LLMs like ChatGPT. If you keep scaling them up, they eventually become worse. This inverse scaling leads to them becoming actively bad: (https://youtube.com/watch?v=viJt_DXTfwA&t=1705s ) (When @OpenAI were developing the highly anticipated ChatGPT-4, it seems they may have hit this problem already. Because instead of scaling up their training sets, like previous iterations of their model, leaks indicate that ChatGPT-4 may actually be 8 x ChatGPT3 models tethered together (https://thealgorithmicbridge.substack.com/p/gpt-4s-secret-has-been-revealed ), explaining why it was delayed and why the dataset size was not revealed.) But even if you believe they’ll find a way around that problem, there’s still plenty of others waiting for us. What if I told you that language model technology isn’t actually new? What if I told you it’s largely the same as it was in the 1980s, but the only thing that’s changed is the transformer technology, allowing for more efficient training, and the sheer size of the training data: The public internet. Yes, the thing that gives ChatGPT (and other LLMs) their “magic” is the fact that the internet exists now and can be scraped. (After it’s being manually catalogued by hundreds of thousands of foreign workers, of course (https://theglobeandmail.com/business/article-ai-data-gig-workers/ ).) 300 billions words from the internet were used to train ChatGPT-4. It’s the scale of this training dataset that allows it to sound so human and knowledgeable. There’s nothing else like it. Not only are major companies preventing their content from being used in future AI training datasets (https://deadline.com/2023/10/bbc-will-block-chatgpt-from-scraping-its-content-1235566868/ ) but there’s a lingering question on whether or not it was even legal for them to use their original datasets in the first place (https://theverge.com/2023/9/11/23869145/writers-sue-openai-chatgpt-copyright-claims ). But worse than that, the internet is increasingly being polluted with error-ridden AI generated content. So much so that it’s infecting search results (https://wired.com/story/fast-forward-chatbot-hallucinations-are-poisoning-web-search/ )…”

Of course they do. We keep coming back to the same ‘government wants’ at every opportunity. Would I be required to own a smartphone in order to drive a car?

https://www.pogowasright.org/the-tsa-wants-to-put-a-government-tracking-app-on-your-smartphone/

The TSA wants to put a government tracking app on your smartphone

Edward Hasbrouck writes:

Today the Identity Project submitted our comments to the Transportation Security Administration (TSA) on the TSA’s proposed rules for “mobile driver’s licenses”.

The term “mobile driver’s license” is highly misleading. The model Electronic Credential Act drafted by the American Association of Motor Vehicle Administrators (AAMVA) to authorize the issuance of these digital credentials and installation (“provisioning”) of government-provided identification and tracking apps on individual’s smartphones provides that, “The Electronic Credential Holder shall be required to have their Physical Credential on their person while operating a motor vehicle.”

So the purpose of “mobile driver’s licenses” isn’t actually licensing of motor vehicle operators, as one might naively assume from the name. Rather, the purpose of the “mobile drivers license” scheme is to create a national digital ID, according to standards controlled by the TSA, AAMVA, and other private parties, to be issued by state motor vehicle agencies but intended for use as an all-purpose government identifier linked to a smartphone and used for purposes unrelated to motor vehicles.

Sunday, October 15, 2023

Any actor can play…

https://www.washingtonpost.com/technology/2023/10/14/propaganda-misinformation-israel-hamas-war-social-media/

A flood of misinformation shapes views of Israel-Gaza conflict

A WhatsApp voice memo purporting to have insider information ricocheted across hundreds of group chats in Israel early on Monday. The Israeli army was planning for another “battle like we’ve never experienced before,” the anonymous woman said in Hebrew, warning that people should prepare to lose access to food, water and internet service for a week.

Across the country, Israelis raced to the banks and to the grocery stores, anticipating another attack. But the message, the army clarified hours later on X, turned out to be a falsehood.

One week into the war between Israel and Gaza, social media is inducing a fog of war surpassing previous clashes in the region — one that’s shaping how panicked citizens and a global public view the conflict.

(Related) Perhaps there was insufficient data in the AI’s library to match the photo?

https://www.404media.co/ai-images-detectors-are-being-used-to-discredit-the-real-horrors-of-war/

AI Images Detectors Are Being Used to Discredit the Real Horrors of War

A free AI image detector that's been covered in the New York Times and the Wall Street Journal is currently identifying a photograph of what Israel says is a burnt corpse of a baby killed in Hamas’s recent attack on Israel as being generated by AI.

However, the image does not show any signs it was created by AI, according to Hany Farid, a professor at UC Berkeley and one of the world’s leading experts on digitally manipulated images.

Perspective.

https://www.zbw.eu/econis-archiv/handle/11159/605045

Disrupting Creativity: Copyright Law in the Age of Generative Artificial Intelligence

Very recently, due largely to breakthroughs in deep learning technologies, AI has begun stepping into the shoes of human content generators and making valuable creative works at scale. Before the end of the decade, a significant amount of art, literature, music, software, and web content will likely be created by AI rather than traditional human authors. Yet the law, as it has so often historically, lags this technological evolution by prohibiting copyright protection for AI-generated works. The predominant narrative holds that even if AI can automate creativity, that this activity is not the right sort of thing to protect, and that protection could even harm human artists. AI-generated works challenge beliefs about human exceptionalism and the normative foundations of copyright law, which until now have offered something for everyone. Copyright can be about ethics and authors and protecting the sweat of a brow and personality rights. Copyright can also be about the public interest and offering incentives to create and disseminate content. But copyright cannot have it all with AI authors—there is valuable output being generated, but by authors with no interests to protect. This Article argues that American copyright law is, and has been traditionally, primarily about benefiting the public interest rather than benefiting authors directly. As a result, AI-generated works are precisely the sort of thing the system was designed to protect. Protection will encourage people to develop and use creative AI which will result in the production and dissemination of new works. Taken further, attributing authorship to AI when an AI has functionally done the work of a traditional author will promote transparency, efficient allocations of rights, and even counterintuitively protect human authors. AI-generated works also promise to radically impact other fundamental tenets of copyright law such as infringement, protection of style, and fair use. How the law should respond to AI activity has lessons more broadly for thinking about what rules should apply to people, machines, and other sorts of artificial authors.

Perspective.

https://www.geekwire.com/2023/robots-ai-and-the-future-of-labor-an-economic-opportunity-way-bigger-than-the-steam-engine/

Robots, AI, and the future of labor: An economic opportunity ‘way bigger than the steam engine’

The global conversation about robots and the workforce has shifted substantially in recent years, from widespread concerns about robots taking jobs to growing questions about how quickly they can fill gaps in the labor market.

One of the ventures at the forefront of this quest is Sanctuary AI. It’s a Vancouver, B.C.-based company that has raised more than $100 million Canadian dollars to pursue its vision for labor as a service. Sanctuary makes a 5-foot, 7-inch general-purpose humanoid robot called Phoenix, powered by an AI system called Carbon.

Perspective. Granted, AI will be faster than humans to see patterns in data. AI will likely find all the patterns in the data. Who selects the data?

https://www.psychologytoday.com/gb/blog/tech-happy-life/202310/the-ai-domino-effect-how-ai-will-soon-outsmart-us-all

The Domino Effect: How AI Will Soon Outsmart Us All

Artificial intelligence (AI) is a civilization-altering technology that is already changing our world in profound ways.

… One thing capitalism is good at is making things better. We merely have to look back to our history of various technologies to see proof of how we improve them—rockets, televisions, video games, laptop computers, phones, etc. There is a powerful, profit-based incentive within our capitalist system to overcome any technical hurdles that stifle technological innovation and evolution. Since there are profits to be made, it's 100 percent guaranteed that capitalism will make AI much better than it is now. Importantly, "better" does not necessarily mean "good."

Saturday, October 14, 2023

Changing how your company works requires human thought, not AI suggestions.

https://www.cio.com/article/655669/business-ai-will-change-the-way-businesses-are-run.html

Business AI will change the way businesses are run

Less than a year after most CIOs and business leaders even heard the expression “generative artificial intelligence,” for the first time, this technology has set off a wave of innovation that will dramatically change how businesses are run.

However, we at SAP are not entering this race as newcomers. In fact, we have been at the forefront of embedding AI into our solutions for years. And as part of this, we recently launched SAP Joule, a generative AI assistant that will be embedded throughout the SAP cloud portfolio.

Joule will help our customers achieve business results faster by enabling them to access insights that are relevant for their business through natural conversation. Simply by asking a question in plain language, our customers will get smart answers drawn from a pool of data from across the SAP portfolio and third-party sources. [Can you control these sources? And what do they think about you using their data? Bob] Joule will continuously deliver new insights that get even more intelligent over time.

Friday, October 13, 2023

How many similar ‘fakes’ are getting by the judges? (This is the area of law you specialize in and you didn’t check citations you have never seen before?)

https://laist.com/news/housing-homelessness/dennis-block-chatgpt-artificial-intelligence-ai-eviction-court-los-angeles-lawyer-sanction-housing-tenant-landlord

This Prolific LA Eviction Law Firm Was Caught Faking Cases In Court. Did They Misuse AI?

Dennis P. Block and Associates, which describes itself as California’s “leading eviction law firm,” was recently sanctioned by an L.A. County Superior Court judge over a court filing the judge found contained fake case law.
Block’s firm was ordered to pay $999 over the violation. That’s $1 below the threshold that would have required the firm to report the sanction to the state bar for further investigation and possible disciplinary action.

Imagine hundreds (thousands?) of these things overhead. And you thought we weren’t ready for driverless cars…

https://www.cnbc.com/2023/10/13/china-lets-ehang-operate-fully-autonomous-passenger-carrying-air-taxis.html

China gives Ehang the first industry approval for fully autonomous, passenger-carrying air taxis

… U.S.-listed Ehang claims it’s the first in the world to get such a certificate, which allows it to fly passenger-carrying autonomous electric vertical take-off and landing (eVTOL) aircraft in China.

The certificate will also significantly simplify the company’s ability to get similar certificates for commercial operation in the U.S., Europe and Southeast Asia, CEO Huazhi Hu told CNBC in a video conference interview.

“Next year we should start to expand overseas,” he said, noting those regulators still need to establish a process for mutual regulation of the Chinese airworthiness certification. That’s according to a CNBC translation of his Mandarin-language remarks.

… A significant difference between self-driving taxis and self-piloting drones is that while cars on the road must make turns at intersections, a drone flight is between two points in the air, Ehang’s CEO said.

Read this one with a grain of salt. I think the doom is overdone. Will some people be left behind? Sure. Some people haven’t learned to use computers yet.

https://news.harvard.edu/gazette/story/2023/10/a-tech-warning-ai-is-coming-fast-and-its-going-to-be-rough-ride/

A tech warning: AI is coming fast and it’s going to be rough ride

… “It’s important for everybody to understand how fast this is going to change,” said Eric Schmidt, the former CEO and chairman of Google, during a conversation Wednesday evening with Graham Allison, Douglas Dillon Professor of Government at Harvard Kennedy School about what’s just over the horizon in AI. “It’s going to happen so fast. People are not going to adapt.”

The pace of improvements is picking up. The high cost of training AI models, which is time consuming, is coming down very quickly, and output quality will be far more accurate and fresher than it is today, said Schmidt.

But “the negatives are quite profound,” he added. AI firms still have no solutions for issues around algorithmic bias or attribution, or for copyright disputes now in litigation over the use of writing, books, images, film, and artworks in AI model training. Many other as yet unforeseen legal, ethical, and cultural questions are expected to arise across all kinds of military, medical, educational, and manufacturing uses.

Most concerning are the “extreme risks” of AI being used to enable massive loss of life if the four firms at the forefront of this innovation, OpenAI, Google, Microsoft, and Anthropic, are not constrained by guardrails and their financial incentives are “not aligned with human values,” said Schmidt, who served as executive chairman of Alphabet, Google’s parent company, from 2015 to 2018, and as technical adviser from 2018 to 2020 before leaving altogether.

Thursday, October 12, 2023

Interesting because I think Adam Smith is a good model for how AI works. Smith studied lists of prices and inventories and found relations between supply and demand. AI does the same things with less structured data (and much faster). What AI can not do is produce new concepts, like Einstein did in his Special Theory of Relativity. (There were no references to prior work in that paper.)

https://www.cnn.com/2023/10/12/economy/ai-impact-on-economists-jobs/index.html

How genAI is revolutionizing the field of economics

Anton Korinek, an economics professor at the University of Virginia, tells the students he advises nowadays that they should really begin to master a flourishing technology expected to transform the field of economics. That technology is generative artificial intelligence, or “genAI” for short.

… Korinek expects it to “revolutionize research,” according to a paper he wrote that was accepted for publication by the Journal of Economic Literature.

“It’s a powerful technology and if you use it, you can solve economic problems that we face as a society, better and more productively. That’s what research is all about,” Korinek told CNN in an interview.

It seems that everyone – even kids – have access to the war fighting tools Hamas found so useful.

https://www.politico.com/newsletters/digital-future-daily/2023/10/10/we-just-saw-the-future-of-war-00120788

We just saw the future of war

When Hamas militants shocked the world last weekend by launching the biggest and most violent attack on Israel in decades, it was almost equally shocking how they did it.

Hamas blasted through a super-high-tech, $1 billion security system on the Gaza border using little more than bulldozers, paragliders and a 2G cellular network, a remarkable upending of the two sides’ tech dynamic — as POLITICO’s Daniella Cheslow outlined in striking detail this morning.

… “Technology is changing warfare, but it isn’t necessarily changing it in the ways that most techno-optimists think it will,” Cronin said. “Because technologies are so accessible, you’ve got… groups like Hamas able to use everything from drones, to social media, to low-tech clusters of technology both high and low that can have an enormous impact.”

For Hamas, that took the form of staying off smartphones and preparing its propaganda in advance, as well as overwhelming the Israeli border so rapidly that its drone surveillance system failed. Cronin characterizes three key areas where lower-tech actors can, and do, overwhelm their counterparts: The democratization of media technology; the increase of physical reach allowed by cheap drones and rocketry; and systems integration, or the ability to communicate effectively within the group.

Wednesday, October 11, 2023

I’ve decided not to answer that question.

https://bigthink.com/the-future/free-will-required-true-artificial-general-intelligence/

Why free will is required for true artificial intelligence

Artificial general intelligence will not arise in systems that only passively receive data. They need to be able to act back on the world.

… Recent years have witnessed stunning advances in other areas like image recognition, text prediction, speech recognition, and language translation. These were achieved mainly due to the development and application of deep learning, inspired by the massively parallel, multilevel architecture of the cerebral cortex. This approach is tailor-made for learning the statistical regularities in masses and masses of training data. The trained neural networks can then abstract higher-order patterns; for example, recognizing types of objects in images. Or they can predict what patterns will be most likely in new instances of similar data, as in the autocompletion of text messages or the prediction of the three-dimensional structures of proteins.

… However, even the most sophisticated systems can quickly be flummoxed by the right kind of questioning, the kind that presents novel scenarios not represented in the training data that humans can handle quite easily. Thus, if these systems have any kind of “understanding” — based on the abstraction of statistical patterns in an unimaginably vast set of training data — it does not seem to be the kind that humans have.

My AI claims that it not nuts.

https://www.psychologytoday.com/us/blog/the-digital-self/202310/dare-we-consider-ai-psychology

Dare We Consider AI Psychology?

In the lexicon of human experience, the term 'psychology' has often been reserved for the intricate maze of the human mind. It encapsulates our emotions, behaviors, perceptions, and much more. However, as we stand at the precipice of a technological revolution driven by artificial intelligence, particularly Large Language Models, the concept of psychology undergoes an interesting metamorphosis. Let’s step off the couch and over the line on a journey into AI Psychology.

Jean Baudrillard’s assertion that “The territory no longer precedes the map” draws a parallel to the current conundrum faced by advanced AI models. Are these systems simply mimicking the vast amounts of data they've been trained on, or is there a deeper level of "understanding"? The term 'hallucination,' commonly associated with cognitive anomalies in humans, is now applied to AI. Such anthropomorphic language leads us to question: Are we inadvertently crafting a unique psyche for these machines or is there some undercurrent of processing that gives rise to pathology?

Makes it easier to ask to be deleted. Maybe.

https://www.latimes.com/california/story/2023-10-10/newsom-bill-delete-online-personal-data

Newsom signs bill that would make it easier to delete online personal data

Californians will be able to make a single request asking that data brokers delete their personal information, under a bill Gov. Gavin Newsom signed into law Tuesday.

Senate Bill 362, also known as the Delete Act, directs the California Privacy Protection Agency to create this new tool by January 2026. [Why so long? Bob]

Is there any evidence of this?

https://www.reuters.com/legal/utah-sues-tiktok-over-impact-app-children-2023-10-10/

Utah sues TikTok, claiming app has harmful impact on children

Utah sued Chinese-owned app TikTok on Tuesday, accusing it of harming children by intentionally keeping young users spending unhealthy amounts of time on the short-video sharing platform.

The Utah suit is the latest action challenging the popular app in the United States, with Indiana and Arkansas bringing similar suits.

Last month, a federal judge blocked California from enforcing a law meant to protect children when they use the Internet.

Tuesday, October 10, 2023

War today: anyone can play. Hack the ‘enemy’ perhaps causing confusion or exposing information or wasting time. All from your favorite couch.

https://www.wired.com/story/israel-hamas-war-hacktivism/

Activist Hackers Are Racing Into the Israel-Hamas War—for Both Sides

Since the conflict escalated, hackers have targeted dozens of government websites and media outlets with defacements and DDoS attacks, and attempted to overload targets with junk traffic to bring them down.

Elon Musk’s X Cut Disinformation-Fighting Tool Ahead of Israel-Hamas Conflict

Elon Musk’s X, in the months before conflict erupted in Gaza, ceased utilizing a software tool used to identify organized misinformation now spreading across the platform formerly known as Twitter.

We asked AI to analyze data and reach a conclusion. Now we are asking it to assume its conclusion was incorrect and reach another conclusion. Sounds like the problem that caused HAL to go crazy.

https://bdtechtalks.com/2023/10/09/llm-self-correction-reasoning-failures/

LLMs can’t self-correct in reasoning tasks, DeepMind study finds

Scientists are inventing various strategies to enhance the accuracy and reasoning abilities of large language models (LLM ) such as retrieval augmentation and chain-of-thought reasoning.

Among these, “self-correction”—a technique where an LLM refines its own responses—has gained significant traction, demonstrating efficacy across numerous applications. However, the mechanics behind its success remain elusive.

A recent study conducted by Google DeepMind in collaboration with the University of Illinois at Urbana-Champaign reveals that LLMs often falter when self-correcting their responses without external feedback. In fact, the study suggests that self-correction can sometimes impair the performance of these models, challenging the prevailing understanding of this popular technique.

An alternative view…

https://www.msn.com/en-xl/news/other/high-schools-in-denmark-are-embracing-chatgpt-as-a-teaching-tool-rather-than-shunning-it/ar-AA1hVLFs

High schools in Denmark are embracing ChatGPT as a teaching tool rather than shunning it

… "My experience was that the students would use it without any kind of thought, and in that way, it becomes an obstacle to learning, and learning is the whole project here," said Pedersen.

"But if we could change the way they use it so that it becomes a tool for learning, then we would have won a lot, both in terms of, well, giving the students a new tool for learning, but also in terms of the relationship with the students," she added.

"Because if we can have the conversation with them about how to use AI, then the whole idea that they can't talk to us about it because it's forbidden goes away.

Monday, October 09, 2023

Heads up:

The Fall Privacy Foundation Seminar is scheduled for Friday, October 27^th.

Recent Developments in State Privacy Laws

*The Colorado Privacy Act

*State Privacy Laws in Colorado compared to states like California and Virginia

Our Panel include:

Corinne O'Doherty, Senior Legislative Aide for Rep. Meg Froelich

Jefferey Riester from Attorney General Phil Weiser's Office

Shelby Dolen from Husch Blackwell

Send comments to the Privacy Foundation at privacyfoundation@law.du.edu and copy John Soma at jsoma@law.du.edu.

Can the law evolve as fast as AI?

https://eprints.ugd.edu.mk/32279/

Civil liability for AI in EU: General remarks

Artificial Intelligence (AI) has become an increasingly prominent and influential technology in modern society, permeating various sectors and significantly impacting the way we live, work, and interact. AI systems possess the ability to analyze vast amounts of data, recognize patterns, and make complex decisions with remarkable speed and accuracy. As a result, AI has facilitated significant improvements in efficiency, productivity, and problem-solving capabilities across various industries. From autonomous vehicles enhancing transportation safety to intelligent virtual assistants streamlining everyday tasks, AI has showcased immense potential in transforming societal functions. Nevertheless, alongside its transformative power, AI presents inherent risks and potential for damage. As AI systems become increasingly autonomous and capable of independent decision-making, questions regarding liability arise. Tort law, with its focus on civil wrongs and compensation for damage, plays a crucial role in determining legal accountability when AI-related damages occur. The complexities surrounding liability in AI are magnified due to the intricate interplay between human agency and machine autonomy, raising challenging legal questions that demand careful consideration.

This contribution aims to present a general overview of the liability regimes currently in place in EU Member States and to determine if they provide for an adequate distribution of all such risks. The starting idea of this research is that such cases in the EU will often have different outcomes due to peculiar features of these legal systems that may play a decisive role, especially in cases involving AI. Mainly, these legal regimes largely attribute liability to human actors, emphasizing concepts such as negligence or intentional misconduct. On the other hand, although there are strict liabilities in place in all European jurisdictions, for the legal theory at present many AI systems do not fall under these regimes, and the victims are left with the sole option of pursuing their claims via fault liability.

How does that work? An obvious question?

https://www.pnas.org/doi/abs/10.1073/pnas.2301842120

Interpretable algorithmic forensics

One of the most troubling trends in criminal investigations is the growing use of “black box” technology, in which law enforcement rely on artificial intelligence (AI) models or algorithms that are either too complex for people to understand or they simply conceal how it functions. In criminal cases, black box systems have proliferated in forensic areas such as DNA mixture interpretation, facial recognition, and recidivism risk assessments. The champions and critics of AI argue, mistakenly, that we face a catch 22: While black box AI is not understandable by people, they assume that it produces more accurate forensic evidence. In this Article, we question this assertion, which has so powerfully affected judges, policymakers, and academics. We describe a mature body of computer science research showing how “glass box” AI—designed to be interpretable—can be more accurate than black box alternatives. Indeed, black box AI performs predictably worse in settings like the criminal system. Debunking the black box performance myth has implications for forensic evidence, constitutional criminal procedure rights, and legislative policy. Absent some compelling—or even credible—government interest in keeping AI as a black box, and given the constitutional rights and public safety interests at stake, we argue that a substantial burden rests on the government to justify black box AI in criminal cases. We conclude by calling for judicial rulings and legislation to safeguard a right to interpretable forensic AI.

It is sounding less and less likely that AI will replace lawyers. Darn.

https://techreg.org/article/view/17979

All Rise for the Honourable Robot Judge?

There is a rich literature on the challenges that AI poses to the legal order. But to what extent might such systems also offer part of the solution? China, which has among the least developed rules to regulate conduct by AI systems, is at the forefront of using that same technology in the courtroom. This is a double-edged sword, however, as its use implies a view of law that is instrumental, with parties to proceedings treated as means rather than ends. That, in turn, raises fundamental questions about the nature of law and authority: at base, whether law is reducible to code that can optimize the human condition, or if it must remain a site of contestation, of politics, and inextricably linked to institutions that are themselves accountable to a public. For many of the questions raised, the rational answer will be sufficient; but for others, what the answer is may be less important than how and why it was reached, and whom an affected population can hold to account for its consequences.

New approach?

https://www.research.unipd.it/handle/11577/3496800

Artificial Intelligence, the Public Space, and the Right to Be Ignored

AI is capable of is occupying, patrolling, and even controlling the physical space. The issue of Artificial Face Recognition is only the most visible aspect of a much broader phenomenon, as computer vision empowers private and public entities to capture a great deal of information that is available in the public sphere in an unprecedented way, challenging how public law conceives public spaces. Despite significant differences, various legal orders are similarly concerned with this phenomenon, which is shifting the boundaries between what is private and what is public. This Chapter i) explains how, because of AI, public spaces are morphing into something new; ii) argues for the protection of anonymity; iii) demonstrates the failures of several contemporary theorizations of public places and that the standard privacy paradigm fails to provide sufficient protection; iv) proposes a new approach to the topic. Drawing from different legal orders, it ultimately argues for a reconceptualization of the public sphere in a way that mitigates the impact of AI on social life.

Saturday, October 07, 2023

Is this “THE” solution or merely “A” solution?

https://arstechnica.com/tech-policy/2023/10/getty-images-built-a-socially-responsible-ai-tool-that-rewards-artists/

Getty Images built a “socially responsible” AI tool that rewards artists

Getty Images CEO Craig Peters told the Verge that he has found a solution to one of AI's biggest copyright problems: creators suing because AI models were trained on their original works without consent or compensation. To prove it's possible for AI makers to respect artists' copyrights, Getty built an AI tool using only licensed data that's designed to reward creators more and more as the tool becomes more popular over time.

… Rather than crawling the web for data to feed its AI model, Getty's tool is trained exclusively on images that Getty owns the rights to, Peters said. The tool was created out of rising demand from Getty Images customers who want access to AI generators that don't carry copyright risks. Peters explained:

A $100 million here, a $100 million there and pretty soon you’re talking real money!

https://www.databreaches.net/data-breach-at-mgm-resorts-expected-to-cost-casino-giant-100-million/

Data breach at MGM Resorts expected to cost casino giant $100 million

Wyatte Grantham-Philips reports:

The data breach last month that MGM Resorts is calling a cyberattack is expected to cost the casino giant more than $100 million, the Las Vegas-based company said.

The incident, which was detected on Sept. 10, led to MGM shutting down some casino and hotel computer systems at properties across the U.S. in efforts to protect data.

Friday, October 06, 2023

Preview of things to come?

https://www.schneier.com/blog/archives/2023/10/deepfake-election-interference-in-slovakia.html

Deepfake Election Interference in Slovakia

Well designed and well timed deepfake or two Slovakian politicians discussing how to rig the election:

Šimečka and Denník N immediately denounced the audio as fake. The fact-checking department of news agency AFP said the audio showed signs of being manipulated using AI. But the recording was posted during a 48-hour moratorium ahead of the polls opening, during which media outlets and politicians are supposed to stay silent. That meant, under Slovakia’s election rules, the post was difficult to widely debunk. And, because the post was audio, it exploited a loophole in Meta’s manipulated-media policy, which dictates only faked videos—where a person has been edited to say words they never said —go against its rules.

I just wrote about this. Countries like Russia and China tend to test their attacks out on smaller countries before unleashing them on larger ones. Consider this a preview to their actions in the US next year.

As I have been warning…

https://www.wired.com/story/fast-forward-chatbot-hallucinations-are-poisoning-web-search/

Chatbot Hallucinations Are Poisoning Web Search

Untruths spouted by chatbots ended up on the web—and Microsoft's Bing search engine served them up as facts. Generative AI could make search harder to trust.

Another problem...

https://www.bespacific.com/evaluating-llms-is-a-minefield/

Evaluating LLMs is a minefield

Evaluating LLMs is a minefield, Arvind Narayanan & Sayash Kapoor. Princeton University. Oct 4, 2023. Authors of the AI Snake Oil book and newsletter

Tools & Techniques.

https://www.bespacific.com/how-to-use-google-bard-2023-a-comprehensive-guide/

How to Use Google Bard (2023): A Comprehensive Guide

Tech Republic: “This is a complete guide on how to use Google Bard. Learn how Google Bard can help you boost your productivity, creativity and more. Bard is Google’s public entry into the highly competitive field of artificial intelligence chatbots, which also includes OpenAI’s ChatGPT. Google intends Bard to be a “creative and helpful collaborator” that people may chat with using natural language. The following guide covers what you need to know as you chat and explore the capabilities of Google Bard.”

Thursday, October 05, 2023

A step in the right direction?

https://techxplore.com/news/2023-10-method-ai.html

Study presents new method for explainable AI

… In their paper "From attribution maps to human-understandable explanations through concept relevance propagation," the researchers present concept relevance propagation (CRP), a new method that can explain individual AI decisions as concepts understandable to humans. The paper has now been published in Nature Machine Intelligence.

… "On the input level, CRP labels which pixels within an image are most relevant for the AI decision process. This is an important step in understanding an AI's decisions, but it doesn't explain the underlying concept of why the AI considers those exact pixels."

For comparison, when humans see a black-and-white striped surface, they don't automatically recognize a zebra. To do so, they also need information such as four legs, hooves, tail, etc. Ultimately, they combine the information of the pixels (black and white) with the concept of animal.

Steps in the wrong direction. Depressing on an otherwise good day…

https://www.bespacific.com/freedom-on-the-net-2023-the-repressive-power-of-artificial-intelligence/

Freedom on the Net 2023 The Repressive Power of Artificial Intelligence

Highlights – Freedom on the Net 2023

Global internet freedom declined for the 13th consecutive year. Digital repression intensified in Iran, home to this year’s worst decline, as authorities shut down internet service, blocked WhatsApp and Instagram, and increased surveillance in a bid to quell antigovernment protests. Myanmar came close to dislodging China as the world’s worst environment for internet freedom, a title the latter country retained for the ninth consecutive year. Conditions worsened in the Philippines as outgoing president Rodrigo Duterte used an antiterrorism law to block news sites that had been critical of his administration. Costa Rica’s status as a champion of internet freedom has been imperiled after the election of a president whose campaign manager hired online trolls to harass several of the country’s largest media outlets.
Attacks on free expression grew more common around the world. In a record 55 of the 70 countries covered by Freedom on the Net, people faced legal repercussions for expressing themselves online, while people were physically assaulted or killed for their online commentary in 41 countries. The most egregious cases occurred in Myanmar and Iran, whose authoritarian regimes carried out death sentences against people convicted of online expression-related crimes. In Belarus and Nicaragua, where protections for internet freedom plummeted during the coverage period, people received draconian prison terms for online speech, a core tactic employed by longtime dictators Alyaksandr Lukashenka and Daniel Ortega in their violent campaigns to stay in power.
Generative artificial intelligence (AI) threatens to supercharge online disinformation campaigns. At least 47 governments deployed commentators to manipulate online discussions in their favor during the coverage period, double the number from a decade ago. Meanwhile, AI-based tools that can generate text, audio, and imagery have quickly grown more sophisticated, accessible, and easy to use, spurring a concerning escalation of these disinformation tactics. Over the past year, the new technology was utilized in at least 16 countries to sow doubt, smear opponents, or influence public debate.
AI has allowed governments to enhance and refine their online censorship. The world’s most technically advanced authoritarian governments have responded to innovations in AI chatbot technology, attempting to ensure that the applications comply with or strengthen their censorship systems. Legal frameworks in at least 21 countries mandate or incentivize digital platforms to deploy machine learning to remove disfavored political, social, and religious speech. AI, however, has not completely displaced older methods of information control. A record 41 governments blocked websites with content that should be protected under free expression standards within international human rights law. Even in more democratic settings, including the United States and Europe, governments considered or actually imposed restrictions on access to prominent websites and social media platforms, an unproductive approach to concerns about foreign interference, disinformation, and online safety.
To protect internet freedom, democracy’s supporters must adapt the lessons learned from past internet governance challenges and apply them to AI. AI can serve as an amplifier of digital repression, making censorship, surveillance, and the creation and spread of disinformation easier, faster, cheaper, and more effective. An overreliance on self-regulation by private companies has left people’s rights exposed to a variety of threats in the digital age, and a shrinking of resources in the tech sector could exacerbate the deficiency. To protect the free and open internet, democratic policymakers—working side by side with civil society experts from around the world—should establish strong human rights–based standards for both state and nonstate actors that develop or deploy AI tools.