Month: March 2023

The Capitol Hill hearings featuring TikTok’s CEO lead off episode 450 of the Cyberlaw Podcast. The CEO handled the endless stream of Congressional accusations and suspicion about as well as could have been expected.  And it did him as little good as a cynic would have expected. Jim Dempsey and Mark MacCarthy think Congress is moving toward action on Chinese IT products – probably in the form of the bipartisan Restricting the Emergence of Security Threats that Risk Information and Communications Technology (RESTRICT) Act. But passing legislation and actually doing something about China’s IT successes are two very different things.

The FTC is jumping into the policy arena on cloud services, Mark tells us, and it can’t escape its DNA; it’s dwelling on possible industry concentration and lock-in and not asking much about the national security implications of knocking off a bunch of American cloud providers when the alternatives are largely Chinese cloud providers. The FTC’s myopia means that the administration won’t get as much help as it could from the FTC on cloud security measures. I reissue my standard objection to the FTC’s refusal to follow the FCC’s lead in deferring on national security to executive branch concerns. Mark and I disagree about whether the FTC Act requires the Commission to limit itself to consumer protection.

Jim Dempsey reviews the latest AI releases, including Google’s Bard, which seems to have many of the same hallucination problems as OpenAI’s. Jim and I debate what I consider the wacky and unjustified fascination in the press with catching AI engaging in wrongthink. I believe it’s just a mechanism for justifying the imposition of left-wing values on AI output – which already scores left/libertarian on 14 of 15 standard tests for identifying ideological affiliation. Similarly, I question the effort to stop AI from hallucinating footnotes in support of its erroneous facts. If ever there were a case for a separate AI citechecker, for generative AI correction of AI errors, the fake citation problem seems like a natural.

Speaking of Silicon Valley’s lying problem, Mark reminds us that social media is absolutely immune for false user speech, even after it gets notice that the speech is harmful and false. He reminds us of his thoughtful argument in favor of tweaking section 230 to more closely resemble the notice and action obligations found in the Digital Millennium Copyright Act (DMCA). I argue that the DMCA has not so much solved the incentives for overcensoring speech as it has surrendered to them.

Jim introduces us to an emerging trend in state privacy law: privacy bills that industry supports.  Iowa’s new law is the exemplar; Jim questions whether it will satisfy users in the long run.

I summarize Hachette v. Internet Archive, in which Judge John G. Koeltl delivers a harsh rebuke to internet hippies everywhere, ruling that the Internet Archive violated copyright in its effort to create a digital equivalent to public library lending. The judge’s lesson for the rest of us: You might think fair use is a thing, but it’s not. Get over it.

In quick hits,

• I note that the Cyberlaw Podcast scooped WIRED in covering the GSA’s lies about the security of login.gov and its later effort to justify those lies by invoking “equity” – currently replacing patriotism as the last resort of scoundrels.
• And I offer a brief, nostalgic requiem for Toshiba, which is being broken up for scrap by what’s left of Japan Inc. Thirty years ago, Toshiba was treated on the Hill like Huawei is today – a scary and unstoppable competitor who threatened the American way of life.  Now, not so much.

Download 450th Episode (mp3)

You can subscribe to The Cyberlaw Podcast using iTunes, Google Play, Spotify, Pocket Casts, or our RSS feed. As always, The Cyberlaw Podcast is open to feedback. Be sure to engage with @stewartbaker on Twitter. Send your questions, comments, and suggestions for topics or interviewees to CyberlawPodcast@steptoe.com. Remember: If your suggested guest appears on the show, we will send you a highly coveted Cyberlaw Podcast mug! The views expressed in this podcast are those of the speakers and do not reflect the opinions of their institutions, clients, friends, families, or pets.

The post China in the Bull Shop appeared first on Reason.com.

from Latest https://ift.tt/spzIeh1
via IFTTT

When hideous murders committed with guns make national news, as in today’s Nashville school shooting, it is a natural reaction on many people’s parts to call passionately for just getting more guns out of more people’s hands, by any means necessary.

Those concerned with civil liberties and police abuse of the innocent have long worried about the effects of privileging mere gun possession arrests. A new study from the Marshall Project highlights some of these problems, via examining data and interviewing arrestees from Chicago’s attempts to manage its violence problem with more gun possession arrests.

Chicago police seem to have adopted a pattern of pretextual stops—looking for any excuse to pull someone over and interact with them—in order, they hope, to find a gun without a proper permit to own or license to carry. (Getting both those things can be quite expensive in terms of both time and money, disproportionately affecting the poor living in violent neighborhoods who most strongly feel the pull toward having a gun for self-defense.) Gun law enforcement in Chicago, the Marshall Project finds, “overwhelmingly focuses on possession crimes — not use.”

The heart of the study’s findings:

From 2010 to 2022, the police made more than 38,000 arrests for illegal gun possession. These arrests — almost always a felony — doubled during this timeframe. While illegal possession is the most serious offense in most of the cases we analyzed, the charges often bear misleading names that imply violence, like “aggravated unlawful use of a weapon.”

Police continue to insist that every gun out of someone’s hands, even if it involves arresting that person and increasing mass incarceration of the nonviolent and highly harming that person’s ability to work and rent in the future, is all for the good; after all, that gun won’t be used by that person to harm anyone. (It is almost always the case that those guns would not have unjustly harmed anyone even minus the arrest and confiscation.)

But there is no particular evidence that harassing or arresting those who have harmed no one has “substantially reduced shootings in Chicago. In fact, as possession arrests skyrocketed, shootings increased, but the percentage of shooting victims where someone was arrested in their case declined,” that latter point a strong indication that mere possession arrests are not a sensible high priority for police for whom gun violence should be their only concern when it comes to guns, not paperwork violations. But the Project found that “nearly 60% of the 31,000 new felony cases pursued by [the state’s attorney of Cook County, Illinois] in the past three years were for illegal gun possession; roughly 4% were for homicides.”

Among the gun arrests the Marshall Project were able to analyze from raw data, “Most people [arrested] had no arrest warrants out, nor were they on supervised release, probation or suspected of being in a gang. In most of the incidents we analyzed, police were not responding to 911 calls about a person with a gun.”

The fight against mere possession becomes a generalized excuse for harassment of citizens: “In arrests where possession was the most severe charge…we found that more than 7 in 10 began with a simple traffic violation. After this initial stop, police often used some other justification for a search. Officers often did this by citing the smell of marijuana. Although Illinois legalized cannabis in 2020, smoking while driving is still prohibited.” And, “in a third of the stops, we found the person arrested had their gun owner’s permit, but not the license that allowed carrying the loaded gun in public.”

The enforcement is hugely disproportionate racially. “Although Black people comprise less than a third of the city’s population, they were more than 8 in 10 of those arrested for unlawful possession in the timeframe we reviewed.” Convictions lead to a year or more in prison, and merely the arrest can mess up the life of someone who has harmed no one via “damning criminal records, time on probation, job loss, legal fees and car impoundments.”

As Reason‘s C.J. Ciaramella has previously reported, the overzealous search for guns owned without proper paperwork has led not just to pretextual traffic stops that ruin innocent people’s lives but to violent raids on homes, often not even the homes where the alleged illegal guns were supposed to be.

As the Marshall Project (who although close-focused on Chicago in this study “identified several other cities with similar trends,” including Houston, New York, Cleveland, Baltimore, and Memphis) quotes one of the victims of gun possession arrests they interviewed, “I’m scared for my life — and I gotta go to prison because I fear for my life, for my family’s safety? Because we’re not fortunate enough to live someplace else?”

The post The Problems With Just Getting Guns Out of People's Hands as a Solution to Gun Violence appeared first on Reason.com.

from Latest https://ift.tt/OS02pIK
via IFTTT

In this week’s The Reason Roundtable, editors Matt Welch, Katherine Mangu-Ward, Nick Gillespie, and special guest Elizabeth Nolan Brown unpack the ubiquitous sense that politicians of every stripe have abandoned a commitment to free expression. They also examine the fast evolution of artificial intelligence chatbots like ChatGPT.

0:42: Politicians choose the culture war over the First Amendment

20:04: Artificial intelligence and large language model (LLM) chatbots like ChatGPT

36:13: Weekly Listener Question

44:27: This week’s cultural recommendations

Mentioned in this podcast:

“Congress Asks Is TikTok Really ‘An Extension of’ the Chinese Communist Party?” by Elizabeth Nolan Brown

“TikTok Is Too Popular To Ban,” by Elizabeth Nolan Brown

“Utah Law Gives Parents Full Access to Teens’ Social Media,” by Elizabeth Nolan Brown

“Florida’s War on Drag Targets Theater’s Liquor License,” by Scott Shackford

“Welcoming Our New Chatbot Overlords,” by Ronald Bailey

“Maybe A.I. Will Be a Threat—To Governments,” by Peter Suderman

“The Luddites’ Veto,” by Ronald Bailey

“Artificial Intelligence Will Change Jobs—For the Better,” by Jordan McGillis

“The Robot Revolution Is Here,” by Katherine Mangu-Ward

“The Earl Weaver Case for Rand Paul’s Libertarianism,” by Matt Welch

“Rand Paul Tries (Again!) To Make It Harder for Police To Take Your Stuff,” by Scott Shackford

Send your questions to roundtable@reason.com. Be sure to include your social media handle and the correct pronunciation of your name.

Today’s sponsor:

• Getting to know yourself can be a lifelong process, especially because we’re always growing and changing. Therapy is all about deepening your self-awareness and understanding. Because sometimes we don’t know what we want or why we react the way we do until we talk through things. BetterHelp connects you with a licensed therapist who can take you on that journey of self-discovery from wherever you are. If you’re thinking of starting therapy, give BetterHelp a try. It’s entirely online. Designed to be convenient, flexible, and suited to your schedule. Just fill out a brief questionnaire to get matched with a licensed therapist, and switch therapists anytime for no additional charge. Discover your potential with BetterHelp. Visit BetterHelp.com/roundtable today to get 10 percent off your first month.

Audio production by Ian Keyser

Assistant production by Hunt Beaty

Music: “Angeline,” by The Brothers Steve

The post Now Is the Best Time To Embrace Artificial Intelligence appeared first on Reason.com.

from Latest https://ift.tt/rKYaE6f
via IFTTT

I am not one to disagree with Eugene about an area within his area of expertise, but I confess I have a different reaction to his view, expressed in his draft paper, that AI program outputs “would reasonably appear to state or imply assertions of objective fact.” Take OpenAI’s ChatGPT.  Eugene argues in his draft at page 8 that OpenAI’s business model is premised on ChatGPT outputs being factually correct:

OpenAI has touted ChatGPT as a reliable source of assertions of fact, not just as a source of entertaining nonsense. . . .  The AI companies’ current and future business models rests entirely on their programs’ credibility for producing reasonable accurate summaries of the facts. When OpenAI helps promote ChatGPT’s ability to get high scores on bar exams or the SAT, it’s similarly trying to get the public to view ChatGPT’s output as reliable. Likewise when its software is incorporated into search engines, or into other applications, presumably precisely because it’s seen as pretty reliable. It can’t then turn around and, in a libel lawsuit, raise a defense that it’s all just Jabberwocky.

Naturally, everyone understands that AI programs aren’t perfect. But everyone understands that newspapers aren’t perfect, either—yet that can’t be enough to give newspapers immunity from defamation liability; likewise for AI programs. And that’s especially so when the output is framed in quite definite language, complete with purported quotes from respected publications.

Here’s my question: Is the reasonable observer test about business models, or is it about what people familiar with the service would think?  Because if the test is about what normal observers would think, it seems to me that no one who tries ChatGPT could think its output is factually accurate.

That’s what makes ChatGPT distinctive and interesting, I think. It combines good writing and ease of language that sounds real, on one hand, with obvious factual inaccuracies, on the other.  It’s all style, no substance. The false claims of fact are an essential characteristic of the ChatGPT user experience, it seems to me.  If you spend five minutes querying it, there’s no way you can miss this.

For example, back in January, I asked ChatGPT to write a bio for me.  This should be easy to do accurately, as there are lots of online bios of me if you just google my name.   ChatGPT’s version was well written, but it had lots and lots of details wrong.

For example, I won't have it be writing my bio any time soon. pic.twitter.com/2b8H01jzxG

— Orin Kerr (@OrinKerr) January 13, 2023

To correct the errors in the ChatGPT output, I joined Berkeley in 2019, not 2018; I didn’t go to Yale Law School; I didn’t clerk for Judge O’Scannlain; I wasn’t an appellate lawyer at DOJ; there is no 2019 edition of my Computer Crime Law casebook, and it certainly wouldn’t be the 2nd edition, as we’re now on the 5th edition already; I’m not a fellow at the American College of Trial Lawyers; and I’ve never to my knowledge been an advisor to the U.S. Sentencing Commission.  (Some would say I’m also not an invaluable asset to the law school community, but let’s stick to the provable facts here, people.)

My sense is that these kinds of factual errors are ubiquitous when using ChatGPT.  It has style, but not substance.  ChatGPT is like the student who didn’t do the reading but has amazing verbal skills; it creates the superficial impression of competence without knowledge.  Maybe that isn’t what OpenAI would want it to be.  But I would think that’s the conclusion a typical user gets pretty quickly from querying ChatGPT.

The post Are AI Program Outputs Reasonably Perceived as Factual? A Response to Eugene appeared first on Reason.com.

from Latest https://ift.tt/qE0PxMY
via IFTTT

Georgetown political philosopher Jason Brennan is one of the world’s leading theorists of democracy, and his new book Democracy: A Guided Tour is a valuable overview of the strengths and weaknesses of democracy, and various arguments on that topic developed by leading political thinkers from ancient Greece to the present day. Brennan has an impressive command of the relevant literature, and he’s a very clear writer. The book is a great resource for anyone interested in debates about democratic theory.

Democracy: A Guide Tour is divided into five sections, each of which considers a leading traditional rationale for democracy: stability, virtue, wisdom in decision-making, liberty, and equality. Each section includes a chapter outlining arguments holding that the democracy effectively promotes the value in question, followed by one outlining reasons for skepticism.

One of the strengths of Brennan’s book is that he carefully avoids the widespread tendency to conflate that which is “democratic” with that which is good. This leaves room for the possibility that increasing democracy might actually cause harm, as opposed to defining it away in advance.

Brennan himself is something of a democracy-skeptic, as evidenced by his previous work. But in Guided Tour, he strives hard to be balanced, and mostly succeeds. Still, it’s hard to come away from the book without getting a sense that democracy falls well short of at least the more expansive claims on its behalf. For example, far from promoting careful deliberation over policy and making effective use of the “wisdom of crowds,” democracy is often plagued by systematic voter ignorance and bias. Far from making citizens more virtuous, democratic political participation often brings out and exacerbates some of the worst tendencies in human nature. And so on.

I myself am also skeptical of many of the benefits claimed for democracy, especially when it comes to issues related to voter ignorance. It’s possible that a reader who comes to the book with the opposite predisposition will view it differently. Either way, the book provides an excellent analysis of a wide range of perspectives on these issues. And it manages to cover a lot of ground in a relatively modest amount of space.

Despite its many strengths, Guided Tour does have a few notable limitations. First, while Brennan carefully considers strengths and weaknesses of democracy in an absolute sense, there is little in the way of comparison between democratic and non-democratic regimes. Brennan discusses a number of thinkers who reject democracy in favor of authoritarianism, but says little about the actual performance of the sorts of regimes they advocate.

Incorporating the latter would make democracy look better on many dimensions. For example, as Brennan explains in Chapter 9, democracy is often a threat to liberty. But authoritarian governments are far worse. For all their many injustices, no democratic government has even come close to the level of oppression perpetrated by communist and fascist regimes. Similarly, a dictator who listens mainly to boot-licking cronies who tell him what he wants to hear will often make even worse cognitive errors than ignorant democratic electorates. Vladimir Putin’s recent disastrous miscalculations may be a case in point.

Brennan is aware of the relative superiority of democratic regimes over authoritarian ones. But it gets little mention in this book, perhaps because the author had to cover so much other ground. Still, I think this issue deserved more attention. The relative failure to consider it may lead readers to (wrongly) dismiss Brennan’s more democracy-skeptical chapters by citing Winston Churchill’s famous statement that “democracy is the worst form of government – except for all the others that have been tried.”

This ties into Brennan’s second notable omission: there is very little discussion of various institutional structures that can potentially mitigate the flaws of democracy, such as federalism, separation of powers, and judicial review. In a previous book, Brennan himself advocates “epistocracy,” giving more power to relatively more knowledgeable voters (see my assessment of the idea here). It’s only briefly mentioned here, as part of a discussion of John Stuart Mill.

Many of these institutional fixes—most obviously, judicial review—are actually constraints on democracy (defined as majoritarian government). Thus, to the extent they are effective, they don’t necessarily vindicate democracy, as such. But there is still a crucial difference between structures that limit democracy and those that dispense with it entirely. If nothing else, the potential benefits of the former undercut Churchill-quoting complacency about the democracy, which is often implicitly premised on the assumption that authoritarianism is the only alternative to giving democratic majorities a virtual blank check to rule as they please.

In sum, Democracy: A Guided Tour is a great overview of various longstanding debates about democracy. But it leaves room for a broader tour that more fully considers non-democratic alternatives to democracy and institutional fixes for various democratic pathologies.

 

The post Jason Brennan's "Democracy: A Guided Tour" appeared first on Reason.com.

from Latest https://ift.tt/QHLJb2u
via IFTTT

At the end of 1980, under the headline “Survey Finds Sharp Rise in Anti-Semitic Incidents,” The New York Times reported that the Anti-Defamation League (ADL) had counted “377 cases of assaults and vandalism…against properties” that year, plus “112 bodily assaults or harassments.” By comparison, the ADL had reported “129 property incidents” in 1979. But Nathan Perlmutter, then the organization’s director, “said that part of the 1980 increase might have reflected improved procedures introduced this year in collecting and evaluating information.”

That sort of caveat is conspicuously missing from this year’s ADL audit and from the Times story about it. “The number of antisemitic incidents in the United States last year was the highest since the Anti-Defamation League began keeping track in 1979,” the Times reports. But there is a difference between “the number of antisemitic incidents in the United States” and the number counted by the ADL, whose annual tally relies on reporting by “victims, law enforcement, the media and partner organizations.”

In addition to actual changes in antisemitic incidents, the ADL’s numbers are apt to be affected by changes in reporting behavior and in the organization’s efforts to collect information. Those factors don’t mean the ADL’s narrative of rising antisemitism in the United States, which is supported by recent survey data, should be dismissed out of hand. But they do complicate the picture in ways that the ADL and the Times fail to acknowledge, especially when it comes to the implicit claim that anti-Jewish bigotry is more prevalent today than it was four decades ago.

Per capita, the ADL counted more than five times as many antisemitic incidents in 2022 as it did in 1980. It does not follow that Americans are five times as likely to hate Jews as they were 43 years ago.

The ADL’s survey data do not support that inference. In 2022, it reports, 20 percent of Americans at least “somewhat” believed “six or more anti-Jewish tropes,” compared to 29 percent in 1964 and 20 percent in 1992. That number has gone up and down over the years, and it rose by 82 percent, from 11 percent to 20 percent, between 2019 and 2022.

During the same period, the number of ADL-reported antisemitic incidents rose by 75 percent, which is consistent with the premise that the organization’s tallies reflect a real trend. But these two sets of numbers do not always track each other so neatly. Between 1992 and 1998, for example, the measure of antisemitic attitudes fell by 40 percent, while the number of incidents that the ADL counted fell by just 7 percent. Between 1998 and 2002, when survey-measured antisemitism rose by 42 percent, the number of reported incidents dropped by 3 percent.

The ADL counted 3,697 antisemitic incidents in 2022, up 36 percent from the year before. A large majority (62 percent) involved “harassment,” which rose by 29 percent. Reports of vandalism, which accounted for 35 percent of the total, rose by 51 percent. Reported assaults, which represented 3 percent of all cases, were up by 26 percent.

The ADL emphasizes that its audit “is not a public opinion poll or an effort to catalog every expression of antisemitism.” And while the “harassment” category includes “individuals or groups” who “were harassed online via antisemitic content in direct messages, on listservs or in social media settings,” the audit “does not attempt to assess the total amount of antisemitism online.”

The ADL says most of the incidents it tallied were reported by victims “via our online form, email or phone message.” The ADL’s researchers also “monitor media reports and other online spaces for credible reports of antisemitic incidents.” In 2021, the ADL “began incorporating reports of antisemitic incidents from other Jewish organizations with whom ADL has established partnerships.” Finally, “many law enforcement partners…share information about antisemitic incidents and criminal activity” with the ADL, and “many of those incidents are included in the Audit as well.”

If you read the Times story, which is headlined “Antisemitic Incidents Reach New High in U.S., Anti-Defamation League Says,” and follow the hyperlink to the ADL’s audit, you will see a button inviting you to “report an incident.” That illustrates one way in which alarming stories about antisemitism might help boost the ADL’s numbers.

More generally, reporting is apt to be influenced by publicity, including the attention attracted by the ADL’s audits and concerns about rising antisemitism. It is plausible that when antisemitic incidents receive a lot of attention, victims are more likely to report them, law enforcement agencies are more likely to take them seriously and keep track of them, and news organizations are more like to cover them. That reaction could create a feedback loop, generating more reports, which in turn generate more publicity.

The ADL’s recent outreach to other Jewish organizations also presumably would have alerted it to incidents it might otherwise have overlooked. What about those “law enforcement partners”? The ADL does not say how many there are or whether the number has increased over time.

While the ADL presents evidence that antisemitic beliefs and conduct have risen in tandem since 2019, in other words, its tallies should not be accepted at face value. Some of the details in the ADL’s report reinforce that point.

“Known white supremacist networks engaged in coordinated efforts to spread antisemitic propaganda, which accounted for 852 incidents in 2022, more than double the 422 incidents in 2021,” the ADL says. “If white supremacist activity had remained the same in 2022 as in 2021, the Audit total would have been 3,267—an increase of 20%, rather than the actual increase of 36%.”

The ADL says three groups—Patriot Front, Goyim Defense League, and White Lives Matter—”were responsible for 93 percent” of the antisemitic propaganda it identified in 2022. The fact that material distributed by a few nutball groups accounted for such a large share of the “harassment” incidents suggests how volatile these numbers can be.

In 2021, for what it’s worth, the FBI counted 396 anti-Jewish crimes, including vandalism, intimidation, and assault. Unlike the 2,717 incidents reported by the ADL that year, all of those cases involved criminal violations. But the FBI’s 2021 number is based on information from just 11,883 of 18,812 law enforcement agencies, down substantially from the years before it revamped its reporting system. Keeping that point in mind, the total for anti-Jewish crimes was 856 in 2016; 1,024 in 2017; 940 in 2018; 1,094 in 2019; and 959 in 2020.

The post Are Americans More Antisemitic Than They Were Four Decades Ago? appeared first on Reason.com.

from Latest https://ift.tt/z62UswP
via IFTTT