Wasatch GC

Regulating AI – Avoiding an “Artificial Intelligence Safety Incident”

Danny Barber — Thu, 29 Aug 2024 19:31:32 +0000

This journal entry provides an update on state efforts to regulate artificial intelligence. Sounds dense, but I will try to keep it straightforward. As always, there are a number of fine details, but to stay entertaining, I will keep it high level. I welcome any more specific questions.

Overall, this entry attempts to distill these new laws down in a way that gives an idea of how state governments view recent advances in the technology.

Since my last entry on artificial intelligence, several states have passed laws regulating aspects of the technology. Each of these laws has their own tilt, with their own distinct objectives.

Notably, the first state in the country to pass such a law was the State of Utah, which in my opinion took a common-sense consumer protection approach.

Others, like SB 1047 (pending California Governor Gavin Newsom’s signature), have taken a more ominous tone.

Starting with the technological advances, for those who haven’t been following along, generative artificial intelligence is “accelerating” (to use a buzz word common in the industry).

“Hyper realistic” videos and images are continuing to blur the line between the authentic and computer generated.

Strides are being made in allowing developers to simply speak their desired code into existence.

Google’s DeepMind can now generate a real-time interactive video game (trained on the classic Doom), with the promise of greatly enhancing immersive VR experiences.

And some people are starting to believe that AI generated images are more real than non-artificial images.

Seems like a lot is happening fast, so how are our state governments going about addressing these advances?

In Utah, our legislators have started off with a fairly common-sense approach. Certain “regulated occupations” (i.e. doctors, lawyers) who deploy generative AI “must prominently” disclose that you are interacting with AI, or AI created content, at the beginning of any communication.

(As someone in a “registered profession”, I’m letting you know now that I create my own content and images.)

Further, “unregulated occupations” who are subject to Utah’s consumer protection laws “must clearly and conspicuously” disclose the use of generative AI when asked by a consumer.

Seems like a good start; simply letting people know that they are in fact not interacting with an actual human might be the best way to train people to recognize increasingly advanced content. That should be even more important when consumers are seeking out specialized knowledge, like medical or legal advice.

In contrast, California’s SB-1047, the “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act”, takes a more ominous tone. Perhaps this is a hint at what is to come with further “acceleration”, from a state who has a high concentration of private computing power.

Overall, the bill is less concerned with consumer protection (in the traditional sense), and focuses more on “artificial intelligence safety incidents”, which I guess is a different type of “consumer protection”.

But what exactly is an “artificial intelligence safety incident?”

According to the legislation, they are events that “demonstrably increase the risk of a critical harm,” whether that is due to a model “autonomously engaging in behavior other than at the request of a user”, stealing or the “escaping” of a model’s weights, “critical failures of technical or administrative controls,” or unauthorized use of a model.

(a non exhaustive list)

So I guess the next question is “what is a critical harm?”

Unsurprisingly, the answer to that question isn’t very pleasant, and includes:

“The creation or use of a chemical, biological, radiological, or nuclear weapon in a manner that results in mass casualties;

Mass casualties…resulting from cyberattacks on critical infrastructure by a model conducting, or providing precise instructions for conducting, a cyberattack or series of cyberattacks on critical infrastructure;

Mass casualties…resulting from an artificial intelligence model engaging in conduct that…acts with limited human oversight, intervention, or supervision… [and] results in death, great bodily injury, property damage, or property loss, and would, if committed by a human, constitute a crime specified in the Penal Code that requires intent, recklessness, or gross negligence….”

Or “other grave harms to public safety and security that are of comparable severity to the harms” described above.

Seems like a pretty dark take on the technology, but given how some industry experts agree with the legislation, it might be a sober one.

So I suppose my third and final question for this entry is “how does California plan to prevent these “artificial intelligence safety incidents?””

Apparently, through several mechanisms, including (but not limited to):

Requiring entities with large compute to “implement reasonable administrative, technical, and physical cybersecurity protections” to prevent misuse “in light of the risks associated with the model;”

“assess whether the covered model is reasonably capable of causing or materially enabling a critical harm;”

implement the capability to promptly enact “a full shutdown” of the model;

create and implement a written “safety and security protocol” and conducting annual reviews of that protocol;

submit to third party auditing of that safety and security protocol;

submit unredacted reports to the California Attorney General verifying compliance with the law (which appear to be immune from Freedom of Information Act requests);

And finally (again, not an exhaustive list), letting the California Attorney General know when an “artificial intelligence safety incident” has occurred.

So what are we to take from all this legislation?

In terms of Utah’s law, it seems there is a concern about consumers being unable to distinguish authentic from artificially generated content. Perhaps some “conspicuous” labeling will help consumers learn to better spot the difference, especially as those boundaries blur.

For California, there seems to be concerns about something far more nefarious, and a desire for the government to play a role in ensuring we avoid a disastrous outcome.

Either way, artificial intelligence will probably continue to accelerate, and in the meantime, Utah and California businesses will need to be aware of their increased obligations related to its use.

The post Regulating AI – Avoiding an “Artificial Intelligence Safety Incident” appeared first on Wasatch GC.

On the Inefficient Economics of University Intellectual Property Licensing

Danny Barber — Thu, 18 Jul 2024 17:34:13 +0000

This journal entry is a critique of the university intellectual property license market, which is based on my experience negotiating many intellectual property licenses on behalf of a variety of different commercial organizations.

In my experience, this market is highly inefficient, due largely in part to university middle men that have obtained perhaps unwarranted control over entrepreneurs and their ideas.

It is not always the case, some universities do a great job.

But more often than not, university intellectual property licensing is a substantial drain on time and capital for ventures attempting to commercialize impactful technology.

So what is this imbalance, and how is it caused?

To explain, we can look to a hypothetical example of the evolution of an aspiring Ph.D. student entrepreneur.

An aspiring Ph.D. student enters their program in order to learn and contribute to the knowledge of a given subject. They choose their institution for a variety of factors; most notable are the existing mentors they will interact with, the facilities that they will have access to, and the notoriety of that institution’s degree.

In exchange, the student not only pays tuition for the opportunity, but is also required to contribute something more intangible: the rights to any intellectual property they develop while affiliated with that institution.

And with any luck, they may develop something during their study that is the seed to an impactful technology that has real world commercial application.

After obtaining their degree, many newly minted Ph.D.’s opt not to follow an academic path, and instead desire to bring their technology to the market. It could be an improvement to a medical device, a piece of software, or other product that could provide immediate tangible benefits to the world.

The problem is, the university now owns that idea, and the entrepreneur must buy it back from them.

Universities often justify this, by pointing to the very same reasons the entrepreneur attended the institution in the first place.

Look at our facilities you had access to, we need to pay for those. Look at the knowledge you had access to, we own that too. Look to the value of your degree, it is derived from those facilities and knowledge.

But where did the actual funds for those facilities and knowledge base come from?

First and foremost, it comes from the tuition the student already paid (and others), which has averaged a 5% inflation rate between 2001 and 2021. That is no surprise to anyone who still carries student loan debt (which universities have no liability for, especially in the case where they provided degree programs with no financial prospects).

And second, it comes from federal grants that the university obtains because of the student researcher.

And now, the entrepreneur is asked to pay again to regain the rights to their own creations.

And these costs often are not nominal. They involve an upfront issue fee for the license, recurring minimum royalty payments to keep the license alive, actual royalty payments when the entrepreneur begins selling a product, and often milestone payments for developmental accomplishments.

(These milestone requirements are often a side effect of the free federal grant monies that paid for the entrepreneur’s research under the Bayh-Dohl act; however the monetization of those milestones is a university creation).

So how do we fix this?

A first answer is to simplify the licensing process. Sure, universities should receive some compensation for documenting the intellectual property and administering it on behalf of the student (who often doesn’t have substantial monetary resources). A flat fee for the rights and repayment of these administrative costs should be sufficient, perhaps with a reasonable return on that upfront investment.

But universities should not be entitled to ongoing royalties, which only divert capital away from the commercial market and the continued development of these inventions.These inventions were largely paid for already, both by tuition and federal grant monies.

Universities should refocus on providing education to allow students to prosper, not playing venture capitalist by holding their student’s ideas captive. We already have many venture capitalists, and they are better at their trade.

This reform will benefit the economy, by streamlining commercialization and allowing entrepreneurs to allocate their capital more efficiently, for example, to things like actually making and marketing their inventions.

The post On the Inefficient Economics of University Intellectual Property Licensing appeared first on Wasatch GC.

Non-Human Intelligence and “Promoting the Progress of Science and Useful Arts”

Danny Barber — Thu, 25 Jan 2024 23:13:41 +0000

Fair warning, this entry is about the relationship between non-human intelligence and the United States intellectual property laws.

Give it a try. I personally find the two topics interesting, or at least more interesting than they initially appear.

I also find them helpful in understanding the process our federal government is currently undergoing in its legislation of artificial intelligence. Another hard topic to grasp, but very worthwhile to keep up on.

And wouldn’t you know it, these topics are also excellent examples for explaining a couple foundational aspects of our intellectual property laws in an approachable way.

Let’s start with the basics. Why do we even have patent and copyright laws in the first place?

At least according to the framers of our Constitution in 1787, it was to “promote the progress of science and useful arts.”

Seems like a good enough reason. And back in 1787, they also thought the best way to accomplish that purpose was “by securing for limited times to Authors and Inventors the exclusive right to their respective writings and discoveries.”

Since that time, two hundred years of law making has made the rules slightly more complex.

So for this introductory entry, let’s keep it simple and only worry about two words.

What do “Author” and “Inventor” mean?

The upfront answer is “Author” and “Inventor” mean only a human being.

And it isn’t that way simply because the issue has never come up.

In fact, the U.S. intellectual property laws have long recognized rights for non-human entities. They can own intellectual property rights, just the same way humans can. They can actually even apply for intellectual property rights on behalf of a human “Inventor” or “Author”.

Apple and Shutterstock acquired a number of their patents and copyrights that way.

And it’s not just the corporations, even the animals are getting into the mix.

That’s a pretty happy monkey right? It’s a wild macaque.

It took that picture. Just like you did with your phone the other day. It did it after it stole the camera from a wildlife photographer. Almost like it knew exactly what it was doing.

And you know what, macaques and a lot of other primates are actually quite smart. Kanzi, a bonobo, could speak with humans.

And if we are getting really honest, macaques can play pong only by thinking about it.

Don’t worry, some humans can do much more than that. We are still on top.

Anyway, if you are a human, you own the copyright rights to any selfie (or photo in general) that you take.

But what about the monkey?

Well, when PETA sued the wildlife photographer to try and get the monkey that same status, they lost. Turns out, congress never “secured” copyright rights to animals. If congress wanted to, they should’ve said so.

The same is true for works produced by another type of non-human intelligence: generative AI.

It’s not that the “writings and discoveries” of generative AI don’t potentially “promote the progress of science and useful arts”, congress just hasn’t “secured” rights for it. So, the AI lawsuits lost for the same reason the monkey did.

Stay tuned, congress might change that soon. The President’s executive order is the start of that process, which isn’t exactly a rapid one.

But I suspect there is a lot of interest by the more traditional non-human entities (corporations) to change the current set up. Especially the pharmaceutical corporations, whose robots are getting pretty darn good at biochemistry.

And if we start doing that, maybe we need to also reconsider how we treat animal intelligence.

But for now, there are no “secured rights” for generative AI as an “Inventor” and “Author”.

Wonder what those framers would think about all this.

The post Non-Human Intelligence and “Promoting the Progress of Science and Useful Arts” appeared first on Wasatch GC.

Thoughts on generative AI in the legal profession

Danny Barber — Tue, 05 Dec 2023 19:09:33 +0000

Entry #3 - The importance of the Author's voice and the ineffable

December 5, 2023

First, a common trope:

If you were to raise an owl without prior experience,

should you try to learn how to raise an owl first,

or just go find an owl egg and figure it out.

This journal entry is about my ~~owl~~ robot.

Yes, I use a robot. More than one in fact.

In hindsight, it took me too long to realize it in the first place.

We are better at different things, sometimes bad for each other, but mostly have a productive relationship.

As far as I can tell…

But there are some things I will just do myself, because what else am I going to do?

What is it that I do? Good question.

Often when I am trying to explain to someone what I do all day, I start small:

“I type a lot.”

It’s a basic joke, but quite effective.

Far better than:

“I can’t really tell you all that much about what I do, sorry.”

It’s much better to start with a modest joke.

It also isn’t a lie, which we lawyers can’t do remember (unless it’s “bluster”, whatever that is).

I actually do type a lot, sometimes all day, and about all kinds of things.

And I care a lot about how I type. I enjoy the act of typing, especially with music. I enjoy choosing the font, paragraph spacing, nerd stuff.

Overall, I enjoy thinking about how the words are going to appear to the reader.

I also really enjoy reading, especially about interesting words.

Some words I find really interesting are: Proprioception, Consciousness, Percipient, Qualia, and Sonder.

Another patent word that is interesting is “lexicographer” (a person who defines words, sometimes new ones).

I guess I should start saying:

“I type a lot, and I read a lot of words (preferably interesting ones, sometimes new ones).”

I’m also trying to learn more about linguistics.

It seems interesting, I’m just not great at it.

But a robot is helping explain it to me.

As far as I can tell…

At a high level, it seems important because language is a “medium through which thoughts, ideas, and knowledge are expressed and communicated.”

So at this point I guess I should start saying:

“I type a lot, I read a lot of words (preferably interesting ones, sometimes new ones), because it’s how thoughts, ideas, and knowledge are expressed and communicated.”

Much better.

Now, because I like words so much, you would think that I would be really worried about some new word robot that is better at words than I am.

You’d be right. It’s fascinating.

In fact, I am a bit surprised at how fast it happened.

But in hindsight, I guess I’ve been contributing to the creation of these robots since I was a kid.

I got my first internet connected computer in the 90’s.

Since then, I’ve gone from:

watching a song download for a full day on Napster,

to:

streaming a song over cellular to my phone in the middle of the mountains, while simultaneously playing that song over the speakers in my car via Bluetooth, while that same car is (mostly) driving itself, and also while I am having fictitious conversations with a word robot just to see what it will say, while using a different search robot to make sure the word robot isn’t just making things up (or in patent speak – “acting as its own lexicographer”).

And it is interesting how it works.

As far as I can tell.

According to it, it is pre-trained on a lot of words, and from those words, it learned how probable a string of more words would be, in response to any new words I ask it.

Quite the trick. It’s “neural network” computational architecture must really work.

But hey, my neural network works to, and is more complex!

If anything, the words I chose to provide my word robot are just as important as the pretty ones it sends back to me.

Don’t get me wrong, it often does have the “best” words, in the “best” order.

Very informative, if occasionally forgivably inaccurate words.

Its words are impressive.

They’ve improved my words, especially my science words.

Its words would otherwise be a death sentence in student loans.

But you know what?

Sometimes, I don’t like its words.

And that is a big relief – because its words are different than my words.

And you know what else, the systems I have access to also have an obvious design flaw: they are only trained on language, and not trained on the “ineffable” aspects of being a human.

Because at the end of the day, lawyers designed a pretty good Turing test – a jury of your peers.

Convincing a group of twelve strangers of your perspective is hard for humans. I don’t think robots are there yet, if they ever will be.

How do you teach a robot to communicate “ineffable” things like sincerity of belief, empathy of condition, and trustworthiness.

A jury of peers seems to accomplish that feat by looking real people in the eyes.

Maybe we should do that more.

And even though we human attorneys can look people in the eyes, we still have rules to make sure we tell the truth.

Apart from Rule 4.1, another famous one (out of a different rule book) is Rule 11. You’d be surprised at how many rule books there actually are.

For non-lawyers, Rule 11 means that I get in big trouble if the words I say to a court are wrong and I don’t fix them (regardless of if I wrote them, or my word robot did).

Some (New York) attorneys already graciously served as examples for the application of this rule, and have been “sanctioned” for relying on their word robot’s wrong answer.

“Sanction” – a delightfully hypocritical word. In this case, it’s the bad definition.

And that makes sense, because a major goal of the court system is to accurately inform people, so they can make the best decision, regardless of the source of the accurate information.

In the end, I am still grateful to my robots.

One helped me create a website in short work, with little to no experience.

One helped me with my logo.

One keeps me more informed, educated, and on my writing toes (at a fraction the cost of traditional education).

And another has helped eliminate a substantial amount of traditional law firm overhead – because I guess private law firm book libraries were a big thing before robots.

What I hope that means is that quality legal advice becomes more affordable, more informed, for more people.

Both in person, and in text.

The post Thoughts on generative AI in the legal profession appeared first on Wasatch GC.

The Purpose of this Legal Journal

Danny Barber — Thu, 30 Nov 2023 07:00:11 +0000

Entry #1 - Hello World.

November 30, 2023

Welcome to my journal, glad you are here. This first entry offers you a guide into what I am trying to accomplish.

Overall, I am creating this journal to write about topics I’m interested in, to discuss how I think about things as an attorney, and in order to reconnect electronically with the many people I no longer get to interact with on a daily basis (but still have much to learn from).

I’ve learned a lot about the practice of law by simply having normal conversations with normal people, and want to actively cultivate more of those.

Based on my experience so far, I think a persistent theme of these entries will be that our modern life is full of complications arising from competing considerations (and alliterations). Often there is no certainty that any one decision is entirely correct, but one needs to be made regardless.

I will also do my best to intermix art and photography that I reflect with for each entry, which I believe helps humanize complex decisions.

Don’t worry, this journal is not strictly legal in nature. While I have no choice but to write with the voice of a patent attorney (yes, I am the one writing these, not my GPT), I will do my best to make my entries interesting, accessible, and not too long.

In any event, I hope you continue to read, and you find at least some of these entries helpful in your own pursuits. I will try to author one every week or so, no promises, it depends on how much real work I have to do.

…and yes, I welcome (preferably constructive) commentary about any of my entries, as well as questions readers want me to investigate privately on their behalf.

However, I cannot guarantee that I will have the relevant skills to provide meaningful responses (which is a requirement of my profession). But I will do my best to help where I can, or connect you with someone who would be able to.

The post The Purpose of this Legal Journal appeared first on Wasatch GC.

Thoughts On Advertising My Law Practice

Danny Barber — Thu, 30 Nov 2023 07:00:00 +0000

Entry #2 - Warning! An "Attorney Advertising"

November 30, 2023

This journal is an advertisement for my law practice (if that wasn’t obvious).

I hope it works.

I also hope it demonstrates how strange advertising a law practice can be.

Put simply, advertising as an attorney is complicated…

…which is probably a good thing?

It turns out, “ambulance chaser” jokes do well at parties. So do a lot of other attorney jokes.

I think the reason why is pretty obvious, we all have an understanding of how a “good attorney” should act, and it is funny when one doesn’t act the way we think they should. (“Things are always looking up for Ol Gil!” – The Simpsons).

What most non-laywers don’t realize is that we attorneys, as an entire profession, have imposed rules upon ourselves that reflect the ways we think “good attorneys” should act. That way, no more new jokes get made.

So it should be no surprise that there are some good rules governing how we as attorneys can advertise ourselves (you know, because of “ambulance chasers”).

After all, shouldn’t lawyers be more interested in genuinely helping people, rather than just monetizing their misfortune?

The first and most obvious rule governing how I can advertise is that I cannot make “a false or misleading communication” about who I am, what I am capable of doing, and what I have achieved. Makes sense right?

Second, any advertisement I make must not use an actor to portray myself, a client, or depict fictionalized events, unless I tell you about it (this one won’t).

I also have to tell you where the ad came from, and how to contact the firm who paid for it (this one came from me, you can contact me on this site).

Last, while it is difficult to tell you exactly how much any one particular project will cost, if I do it in an advertisement, I have to describe “all relevant charges”, which completely depends on the specifics of your situation, and is why no one does it.

These all make sense for advertisements…

But what about if I just go out and talk to people I think I could help?

Yes, there are rules about that too.

Recently, I took my first business trip on behalf of my new law practice.

I went to the 2023 meeting of the Society for Neurosciences, which consists of a collection of varied people who are really interested in figuring out how our brains work.

There are also some people trying to monetize that understanding, which is more difficult than you’d think.

One other benefit was that this convention happened to be in Washington D.C., a city that is often associated (though not always) with some very “good attorneys.”

It also made sense for my background. I have been learning about neuroscience for a while now (15 years), and I’ve worked in the industry for almost just as long.

You would think I should just go talk to people with shared interests at this convention, tell them about myself and my skills, and perhaps I would be able to help them.

But how is that really any different than being an “ambulance chaser?”

As lawyers, we have decided that “it’s complicated.” In Utah, we have agreed not to “solicit professional employment” in person with the “significant motive of obtaining monetary gain”, unless it is from:

Another lawyer,
A family member,
Or someone we already have a close personal or professional relationship with.

So how exactly am I supposed to go advertise myself to the (mostly) strangers at the brain conference?

A good friend helped me with the answer:

it’s not about me, it’s about them.

I should be going to this convention to reconnect with close personal and professional contacts that are also attending. It’s important to still talk with people who you already know and learn about their new challenges.

I should also be going to this conference to learn about my industry, and to challenge myself scientifically by listening to what other people have to say.

In fact, as a “Registered Patent Attorney”, I am ethically obligated to posses “the scientific and technical training necessary to provide valuable service to patent applicants.” (Yes, another rule.)

This makes my “significant motive” not about me or money, but about how my industry is progressing, what new science has come out, what people are worried about, and how I can genuinely help them.

I think that makes sense to anyone in advertising…And it’s actually a pretty a good way to bypass the awkwardness of networking.

In the end, the hard part about creating this journal is that I must be “truthful in statements to others” (Rule 4.1, rules will be a theme), but sometimes the truth is complicated.

It’s also hard to write because I used to read other people’s emails for a living…

Ultimately, there is no rule against you contacting me if you think I can help you understand a complex problem better. But until you do, and we both agree it makes sense, I am not your attorney, and this entire journal is not legal advice directed specifically at you.

I’m just thinking out loud, so you can read about how I think about being a “good attorney” for myself.

The post Thoughts On Advertising My Law Practice appeared first on Wasatch GC.