Patent Trolling to Save the World
post by Double · 2025-01-17T04:13:46.768Z · LW · GW · 0 commentsContents
Activist patent trolls can contribute substantially to every major Effective Altruism priority: Issues with activist patent trolling: What can be done now? None No comments
(Epistemic status: I know next to nothing about patent law, I'm just sharing some thoughts. I would love to be corrected by someone knowledgeable.)
If you think that some technology has a significant chance of ending the world (or having other huge negative externalities that outweigh the benefits), you might wish to postpone that technology. Convincing regulators or pressuring companies not to use a technology are two ways, but could another be to develop the technology first, patent it, and go to court to prevent anyone from using it?
“...patent trolling or patent hoarding is a categorical or pejorative term applied to a person or company that attempts to enforce patent rights against accused infringers far beyond the patent's actual value or contribution to the prior art,[1] often through hardball legal tactics (frivolous litigation, vexatious litigation, strategic lawsuits against public participation (SLAPP), chilling effects, etc.) Patent trolls often do not manufacture products or supply services based upon the patents in question.”-Wikipedia, emphasis added.
A quick search couldn't find any examples of inventors doing this, though there are plenty of inventors who regret their inventions. There's an alternate history where the inventors of vaping or pop up ads realized that their inventions, although technically legal and extremely profitable, would cause huge harm, patented them, and prevented any copycats.
If activist patent trolls can exist, then supporting them can be a hugely beneficial thing to do. By definition, an activist patent troll must:
- In order to invent a valuable technology, they must be among the most knowledgeable in a certain subject. The activist patent trolls for biology must be master biologists, for computer science they must be master computer scientists, etc.
- In order to give up the profits of the technology, they must be extremely altruistic.
Both of these qualities make activist patent trolls among the best decision makers when it comes to the future of technology. I look forward to hearing your thoughts on the political philosophy of ethicotechnotrollocracy.[1]
Activist patent trolls can contribute substantially to every major Effective Altruism priority:
- Animal-welfare activist patent trolls can invent new ways to profitably worsen the lives of animals and patent them. If the horrid inventions powering factory farms like battery cages had been patented by an animal lover and denied the farms for twenty years, then billions of animals would have been better off.
- Global catastrophic risks can be mitigated. For example, if a researcher designed a technique that makes AI more dangerous, then patented it, then that would reduce AI x-risk. (I’ve personally ideated a few ideas that might fall into this category.)
- Inventing new polluting technologies and patenting it can prevent countries in the intersection of “has patent enforcement” and “does not have pollution controls” from using it, improving global health. India falls into this category.
Issues with activist patent trolling:
- There’s a thing called a “Research Exemption” which may prevent activist patents from having much effect on academia. Despite this, if a research direction doesn’t have industry backing and cannot be applied usefully, research usually diminishes. For example, since mirror life [LW · GW] is mostly an academic pursuit and not an industry one, activist patents would not have much effect. On the other hand, if for example Flash Attention was patented by an activist patent troll, then transformers would be much slower to use, AI technology wouldn’t be so profitable, and so there would be fewer researchers working on potentially dangerous capabilities, even if the researchers themselves were allowed to use flash attention.[2]
- Patents only last twenty years, so if an inventor is more than twenty years ahead of her time, then her attempt to slow down the technology by inventing it first backfired. There are large debates about which inventions would have been invented soon anyways even if a particular inventor had not invented it. My (uninformed) position is that most profitable inventions are invented within twenty years of them becoming possible. Each expert, when considering whether to invent something to activist-patent it, should consider how soon someone else would create it.
- With AI technology, labs are secretive and could likely get away with using capabilities-increasing patented technology. Thoughts?
- Patents mostly only apply within countries, so activist patents would have limited scope. Since right now the US is the lead in research and industry, US-based patents can still have significant influence over the use of technology in the rest of the world.
- There’s a sense that activist-inventing-and-patenting is not a neglected issue, since there are countless people trying to make any profitable technologies they can. Even so, I think it is a problem that inventors are not aware of this option, so raising awareness and making it easier to be an activist patent troll could be significantly neglected cause areas.[3]
What can be done now?
- Nonprofit organizations can buy patents from inventors. This would involve the inventor filling out an application to convince the nonprofit of the profitability and negative externalities of the invention. The nonprofit can then go to court if someone infringes the patent. Altruistic inventors would be willing to sell their patent rights to the nonprofit at a discount relative to industry.
- It makes sense that few people have tried to be activist patent trolls, since good people will rarely consider that a career of “design the most harmful yet profitable inventions possible” can actually be a significant force for good. If activist patent trolling is possible, then spreading the word among the people most likely to create harmful inventions can be hugely impactful.
- ^
ethical (ethico), skill/craft/expertise (techno), troll rule (trollocracy)
- ^
Assuming that Flash Attention can be patented. Software patent law is tricky. Is "Flash Attention" an 'abstract idea' (cannot be patented) or 'a product of engineering' (can be patented)?
- ^
Again, assuming it is an option. I’m no expert in patent law.
0 comments
Comments sorted by top scores.