Search the archive.

[Gerald, “Inventors would own more, were it not for noncompliant Bayh-Dole practice,” Research Enterprise, 3/10/23, https://researchenterprise.org/2023/03/10/inventors-would-own-more-were-it-not-for-noncompliant-bayh-dole-practice/, accessed 7/17/24, GDI-JCR]

University ownership is a black hole. The black hole. What’s owned by a university will be used by no one. That’s almost generally true. If a university owns, and insists on trying to find an exclusive licensee, then what are all the other companies to do who don’t need or want exclusivity, or don’t expect to be the exclusive licensee? They will have to design around, ignore, obsolesce, obtain patents that disrupt lines of development or application, and generally undermine that exclusive position to the extent that it represents a threat to competition. You see it, don’t you? An exclusive licensing premise drives everyone else away from university collaboration, away from the inventions a university owns. Makes university research even less relevant to industry, prevents platform and standards development, disrupts early adopters, denies graduating students opportunities to work in industry settings (only at the company, if ever, that gets an exclusive license) with what they have learned in their university research. Out of the blue and into the black. My my, hey, hey.

Will emerging technology cause nuclear war? For more than 70 years, the world has avoided major-power conflict, and many attribute this era of peace to nuclear weapons.1 In situations of mutually assured destruction, neither side has an incentive to launch a nuclear first strike because doing so will only result in self-annihilation. Maintaining secure, second-strike capabilities—the ability to absorb an enemy nuclear attack and respond with a devastating counterattack—is the key to deterrence.2 Recently, analysts have begun to worry, however, that new military technologies may call into question this model of global strategic stability.3

The world is experiencing a fourth industrial revolution (4IR) in which a wave of new and transformative technologies is being developed, including artificial intelligence (AI), additive manufacturing, quantum information technology, hypersonic missiles, biotechnology, and directed energy.4 While these technologies are expected to have profound implications for societies and economies, most are dual use and will also affect national security, including nuclear strategic stability.

[Christopher, Sarah, “The Innovation Wars: America’s Eroding Technological Advantage,” Foreign Affairs 100.2, GDI-JCR]

Since the early days of the Cold War, the United States has led the world in technology. Over the course of the so-called American century, the country conquered space, spearheaded the Internet, and brought the world the iPhone. In recent years, however, China has undertaken an impressive effort to claim the mantle of technological leadership, investing hundreds of billions of dollars in robotics, artiicial intelligence, microelectronics, green energy, and much more. Washington has tended to view Beijing’s massive technology investments primarily in military terms, but defense capabilities are merely one aspect of great-power competition today—little more than table stakes. Beijing is playing a more sophisticated game, using technological innovation as a way of advancing its goals without having to resort to war. Chinese companies are selling 5G wireless infrastructure around the world, harnessing synthetic biology to bolster food supplies, and racing to build smaller and faster microchips, all in a bid to grow China’s power.

[Christopher J. Jr., et al. "The Hidden Cost of University Patents." Berkeley Business Law Journal, vol. 21, no. 1, 2024, pp. 203-249. HeinOnline -tm]

Applying a firm behavior approach to analyzing university activity in the patent space, we have discovered evidence that universities act rationally, if inefficiently. That is, they respond to economic incentives, as any firm would. As expected, universities started patenting innovations produced by their researchers and began to license or sell those patents to businesses.159 They also engage in transactions as a firm would. They created technology transfer offices to manage their patent portfolios and policies. Inevitably, those technology transfer offices became institutional fixtures that had to justify their existence by producing revenue or other activities. Most of what they did was require researchers to ensure the patentability of their research and file as many patent applications as possible. These effects are arguably attributable to the passage of the Bayh-Dole Act.

[Gerald, “Inventors would own more, were it not for noncompliant Bayh-Dole practice,” Research Enterprise, 3/10/23, https://researchenterprise.org/2023/03/10/inventors-would-own-more-were-it-not-for-noncompliant-bayh-dole-practice/, accessed 7/17/24, GDI-JCR]

In my over 20 years in tech transfer, inventions of any use don’t go “into a black hole” if a university fails to demand ownership. Sure, they may not get used–but 95% of patented inventions don’t get used anyway. One should not have a shocked face. But even in this context, the university demand to own everything is an even bigger, blacker black hole.

Following the adoption of Bayh-Dole, university administrators continued to amend their IP policies to expand claims of ownership over inventions created on campus, even claiming rights in the inventions of students, visitors, and other members of the university community.130 Some policies justify these broad ownership claims based on the requirements of the Bayh-Dole Act and other federal laws.131 MIT’s Policy Statements relating to patents provide: Research contracts sponsored by the Federal Government are subject to statutes and regulations under which MIT acquires title in inventions conceived or first reduced to practice in the performance of the research. MIT’s ownership is subject to a nonexclusive license to the government and the requirement that MIT retain title and take effective steps to develop the practi- cal applications of the invention by licensing and other means.132 Pennsylvania State University’s policy declares that its requirement that all members of the university (including adminis- trators, faculty, staff, graduate and undergraduate students, and others) sign an intellectual property agreement arises from the university’s “obligations” under the Bayh-Dole Act.133 However, one cannot lose sight of the fact that the Bayh- Dole Act only applies to federally funded research, not to industry- sponsored research or research that is supported by university funds.134 Nevertheless, it appears that many universities seized upon Bayh-Dole as a justification to claim ownership of all research conducted on campus, rather than limiting their ownership claims to federally funded research.135

National Fire Protection Association, “NFPA Praises Overwhelming Bipartisan Passage of Pro Codes Act,” ASPE Pipeline, 05-06-2024, https://aspe.org/pipeline/nfpa-praises-overwhelming-bipartisan-passage-of-pro-codes-act/

The National Fire Protection Association (NFPA), a self-funded nonprofit that develops over 300 fire, life, and electrical safety standards, praised the bipartisan passage of the Pro Codes Act (H.R.1631) by the House Judiciary Committee. The bill was introduced by Subcommittee Chairman Darrell Issa (R-CA) and Rep. Deborah Ross (D-NC).

Another unchallenged belief of a super AI takeover, with little evidence, is that a super, near-infinite intelligence can quickly solve our major unsolved problems.

Many proponents of an explosion of intelligence expect it will produce an explosion of progress. I call this mythical belief “thinkism.” It’s the fallacy that future levels of progress are only hindered by a lack of thinking power, or intelligence. (I might also note that the belief that thinking is the magic super ingredient to a cure-all is held by a lot of guys who like to think.)

Ben Dickson, “Why AI can’t solve unknown problems,” Venture Beat, 04-02-2021, https://venturebeat.com/ai/why-ai-cant-solve-unknown-problems/

But what everyone agrees on is that current AI systems are a far shot from human intelligence. Humans can explore the world, discover unsolved problems, and think about their solutions. Meanwhile, the AI toolbox continues to grow with algorithms that can perform specific tasks but can’t generalize their capabilities beyond their narrow domains. We have programs that can beat world champions at StarCraft but can’t play a slightly different game at amateur level. We have artificial neural networks that can find signs of breast cancer in mammograms but can’t tell the difference between a cat and a dog. And we have complex language models that can spin thousands of seemingly coherent articles per hour but start to break when you ask them simple logical questions about the world.

*David De Cremer and **Garry Kasparov, “AI Should Augment Human Intelligence, Not Replace It,” Harvard Business Review, 03-18-2021, https://hbr.org/2021/03/ai-should-augment-human-intelligence-not-replace-it

In general, people recognize today’s advanced computers as intelligent because they have the potential to learn and make decisions based on the information they take in. But while we may recognize that ability, it’s a decidedly different type of intelligence what we posses.

Michael Shellenberger, “Why Apocalyptic Claims About Climate Change Are Wrong,” Forbes, 11-25-2019, 

Environmental journalists and advocates have in recent weeks made a number of apocalyptic predictions about the impact of climate change. Bill McKibben suggested climate-driven fires in Australia had made koalas “functionally extinct.” Extinction Rebellion said “Billions will die” and “Life on Earth is dying.” Vice claimed the “collapse of civilization may have already begun.”

Daniel Deudney, “2: Questions, Debates, and Frameworks,” Dark Skies: Space Expansionism, Planetary Geopolitics, and the Ends of Humanity, Oxford University Press 2020, pp. 38.

Is this strongly optimistic view of space and macrodisasters well founded? In opening this question and weighing space ventures as sources of such threats as well as solutions to and escapes from them, this analysis raises the possibility that many significant space activities should be restrained or relinquished on the same grounds that many now believe them to be an unalloyed good, and that human survival may actually hinge on the ability of governance to avoid rather than undertake major space expansions. It is certain that the Earth will become uninhabitable over the very long term of hundreds of millions and billions of years into the future, and that humanity can survive only by leaving the Earth and migrating elsewhere in the cosmos. But these threats are so temporally distant that they can have absolutely no relevance to current human choices and activities. Many of the threats imminent within the next centuries are significantly of human origin, and space expansionists rarely consider the possibility that the patterns of human activity producing these threats might be carried into space along with the spread of humans. Nor do they consider whether regimes to restrain threatening technologies, if established on Earth, can persist if human polities spread across the solar system.

Brian Wang, "Artificial General Intelligence Limitations," NextBigFuture, 6-8-2021, https://www.nextbigfuture.com/2021/06/artificial-general-intelligence-limitations.html

What is Hard and What is Surprisingly Easy

Qinqie He, “The Commercialization of The Domestic AI industry in 2022,” Equal Ocean, 10-5-2022, https://equalocean.com/analysis/2022100518976

Artificial intelligence's growing importance and relevance to technological innovation are undisputed in the digital age. Since 2016, the development of AI has been increasingly embedded in more than 40 countries' national strategies, and governments have started to see artificial intelligence as a nation-defining capability. Under the impact of the Covid-19 pandemic, an increasing number of countries have recognized that AI plays a crucial role in enhancing global competitiveness and have decided to create formal national AI frameworks. Countries that aren't investing in the AI industry risk becoming obsolete.

The Economist, “OpenAI’s legal battles are not putting off customers—yet,” 03-06-2024, https://www.economist.com/business/2024/03/06/openais-legal-battles-are-not-putting-off-customers-yet

So far such legal hurdles are not putting off customers and investors. According to SimilarWeb, a data firm, Openai clocked 1.6bn visits to its mobile app and website in February, five times what Gemini, Google’s rival ai, has managed. Surveys of large companies invariably find Openai and Microsoft to be the most popular purveyors of ai tools. In December Openai was making an annualised revenue of $2bn, up from virtually nothing the year before. Venture capitalists report that the startups they back see its models as best on price and performance. This week it emerged that Temasek, a Singaporean sovereign-wealth fund, wants a stake in the firm.

Matthew Smith, “The New York Times Wants ChatGPT Gone. Nice Try Copyright murkiness is about the only sure thing that lies ahead,” IEEE Spectrum, 01-06-2024, https://spectrum.ieee.org/chatgpt-new-york-times

Whenever it’s resolved, the impact of The New York Times’s suit will be significant. Its failure would allow those looking to train new AI models to proceed as they have in the past with far less concern about future lawsuits. Success wouldn’t entirely halt the training of new AI models, but licensing costs may increase the cost of training beyond the means of all but the most well-funded companies.

Edward Lanquist and Jeremy Ray, “Artificial Intelligence and Copyright Law: The NYT v. OpenAI – Fair Use Implications of Generative AI,” Baker Donelson, 02-05-2024, https://www.bakerdonelson.com/artificial-intelligence-and-copyright-law-the-nyt-v-openai-fair-use-implications-of-generative-ai

While we expect significant clarity from courts over the coming year concerning the application of fair use to generative AI, at the end of the day well-funded industry leaders such as OpenAI will likely win regardless of the outcome. On the one hand, should OpenAI prevail across the board, it will owe nothing for the content used to train its LLM. On the other hand, if OpenAI is forced to license the content used to train its LLM, then such a finding will likely create an economy where only the most well-funded companies will be able to afford the licenses necessary to effectively train their LLMs.

Matthew Smith, “The New York Times Wants ChatGPT Gone. Nice Try Copyright murkiness is about the only sure thing that lies ahead,” IEEE Spectrum, 01-06-2024, https://spectrum.ieee.org/chatgpt-new-york-times

And there’s yet another complication: recent agreements between OpenAI and other publishers, such as Axel Springer and the Associated Press. The exact terms of the deals are unknown, but a press release from OpenAI states its deal with Axel Springer will help the publisher summarize “selected global news.” OpenAI will in turn obtain content from Axel Springer to use for future AI training. OpenAI also reached an agreement with the Associated Press.

Charlotte Tobitt, “Legal experts say OpenAI has ‘case to answer’ in showdown with New York Times,” Press Gazette, 01-11-2024, https://pressgazette.co.uk/media_law/new-york-times-openai-microsoft-ip-experts/

Eziefula believed there are “some noticeable omissions” in how OpenAI has responded so far. “For example, when setting out its position in relation to the US concept of ‘fair use’, OpenAI alludes to the right under EU law that permits Text and Data Mining (TDM). However, OpenAI does not consider the limitations to that right, notably that it only applies to ‘lawfully accessible works’.

Edward Lanquist and Jeremy Ray, “Artificial Intelligence and Copyright Law: The NYT v. OpenAI – Fair Use Implications of Generative AI,” Baker Donelson, 02-05-2024, https://www.bakerdonelson.com/artificial-intelligence-and-copyright-law-the-nyt-v-openai-fair-use-implications-of-generative-ai

The court in The NYT v. OpenAI matter will likely bifurcate its analysis between alleged misuses related to training the LLMs (which, based on the case law set forth below, are likely to be found transformative and fair use) and specific generative AI outputs (with a focus on whether such outputs are substantially similar to specific inputs). Addressing the latter example first, let us assume that an AI company has access to a copyrighted work without a license, uses such work to train their LLM, and then creates a generative AI output that is substantially similar to the copyrighted work. Unless the output is determined to be transformative or meets another requirement of fair use, then such output is likely to be found infringing under the current copyright framework. After all, to prove copyright infringement, the copyright holder merely needs to prove that the alleged infringer has access to the copyrighted work and creates a substantially similar work.