AI and blockchain (in their many forms) are seen as key enabling technologies for the future.
They are both viewed as having the potential to provide world-changing applications which will improve many lives. However, both technologies face an issue with public opinion, which will likely provide an obstacle to them achieving their potential.
Public opinion vs. technical progress
Some of the main barriers associated with the widespread adoption of AI include concerns about a lack of transparency and considerable energy consumption. Rightly or wrongly, these perceived limitations mean that many in the public see AI as much a problem as a solution.
Blockchain also faces challenges with public perception. It has a historical association with cryptocurrency transactions, which (again, rightly or wrongly) are often connected in the public mind with criminal activity. This overlooks any benefits which can be provided, such as those applications which can enhance transparency in data transactions and provide decentralisation, which increases resilience against malicious activities against a blockchain network. Blockchains can also provide these solutions in an increasingly energy-efficient way, following criticisms of earlier implementations for excessive energy use.
Part of the issue with the court of public opinion is that it can often focus on specific, high-profile negative stories, which tend to dominate public discussion. Social media can enable even the most innocuous story to spread quickly and amplify a swirl of unnecessary negativity. This can overshadow the positive impact of a technology, as the desire for outrage and the trust in “groupthink” can often attract more attention than any sort of desire for technical literacy.
Is there a marriage of convenience?
Given that blockchain can provide improved transparency and increasingly energy-efficient applications – and that the absence of these qualities is often seen as being an obstacle to gaining public trust in AI-based solutions – there does seem to be a synergy between the two, which could help both to overcome their challenges with public trust. Specifically, there are several technologies emerging from the development of blockchain which offer:
- Cryptographic security (e.g. homomorphic encryption, multi-party inputs);
- Transparency (e.g. zero-knowledge proofs); and
- Consensus mechanisms (e.g. proof of stake mechanisms and byzantine fault tolerance).
These technologies can improve transparency in AI models by improving trust in computations, which are often offloaded to an untrusted party (which runs the trained model) and assist in showing that the correct steps (in a trained model) have been followed.
Can these technologies be patented?
The application of one technology in a first technical field to a problem in a second field is patentable, provided it solves a technical problem in a non-obvious way. This is irrespective of whether that technical field relates to blockchain or AI, or even if it improves a specific application of either.
Many of the technologies developed for blockchain and AI have a distinctly mathematical character, but this does not mean they would be excluded from patentability. Under the European Patent Convention (EPC), mathematical methods can be protected provided they address a technical problem. This principle has long enabled developments in fields such as signal processing (e.g. applying Fourier transforms and statistical inference) and cryptography (e.g. applying elliptic curves and other abstract techniques) to be protected using patents.
However, it could be said that the application of principles such as homomorphic maps (which generally pops up abstractly in discussions around group theory and topology rather than trustless systems) and Lie group algebra (often used to study high dimensional spaces rather than in the discussion of neural networks) typically delve further into the abstract than what Examiners at patent offices are accustomed to assessing. Nevertheless, provided the application of these advanced mathematical techniques solves a technical problem, the application can still be patented.
Patenting considerations – what should you do?
The patent landscape around blockchain and AI has been well discussed in recent times. Suffice to say, there is a lot of patenting activity when it comes to both blockchain and AI. Many applications do not proceed to grant stage, but many of them are published and so become part of the prior art – this means there is a dense bank of evidence for patent offices to cite when assessing an invention for novelty and/or inventive step.
To improve chances of success in patenting in the blockchain and AI space, and also technologies which apply techniques from one field (e.g. blockchain) into the other (e.g. AI), it is advised to draft patent applications carefully.
Whilst it is desirable to start the patent process with a broad definition of the invention, the patent application should contain plenty of options to enable the claims to be focused more specifically if they need to be. This approach can improve the chances of overcoming objections based on what is already disclosed in the public sphere.
Where an invention relies heavily on mathematics, it is also advised to describe the mathematics carefully and with as much detail as possible. Whilst applications are crafted for the understanding of the skilled person and some knowledge can be assumed, relying on a skilled person to know some of the mathematical steps you believe to be trivial can be problematic. This is because there is the possibility for an inventive step/sufficiency squeeze, where assuming the skilled person has a high level of knowledge, thereby enabling them to fill lots of gaps in the patent specification, can also lead to a higher bar on inventive step, as more knowledge will be taken to be obvious.
Moving forward
AI and blockchain have the potential to assist each other in their respective battles with public opinion. Blockchain can help AI to achieve greater transparency and auditability, whilst AI can help blockchain find more positive, value-driven applications. There is also clear potential for patentability of innovations which merge the two together.
Time will tell how the public’s view of these technologies will change, but their potential to provide world-changing applications will, for now, remain. Even in uncertain times, technological advancement continues, and those who are innovating to address problems with blockchain or AI should seek to protect those technologies in an increasingly crowded patenting landscape.
Our thanks to the Patent Lawyer Annual for publishing the original article. You can read the full version in this PDF (page 16).
