Artificial Intelligence (AI) has been with us for approximately four decades now and most people interact with AI daily and do not realize it. The modern field of AI came into existence in 1956, but the recent accelerated evolution of the technology behind AI and the use of AI is what has increased over the past five-ten years.
The evolution of AI has been divided by researchers and software developers into three stages: the pre-deep-learning era (pre-2010), the deep-learning era (from the early 2010s) and the large-scale developed machine learning era (from 2015). Companies such as Microsoft have been at the forefront of cutting-edge research in AI dating back to the foundation of Microsoft Research (MSR) in 1991and have been integrating these innovative AI technologies into their products such as Windows, Xbox, Microsoft 365, Teams (and MSWord), Azure AI, Power Platform, Dynamics 365, and Microsoft Defender among others[1]. The Google Search experience for instance is powered by hundreds of powerful algorithms. Google continues to train and develop its algorithms over the years to help deliver relevant and accurate search results. RankBrain for example, was developed by Google in 2015 as the first deep-learning system Google AI deployed. RankBrain improved Google’s understanding of how words translate into specific concepts by enabling it to find more information than it previously could and expanding the machine’s understanding of real-world ideas and connections and ranking the best order for “top search” results[2] before the use of search engine optimization became popular.
The recent launches of ChatGPT[3] which interact in a conversational way and Bard[4] which is powered by Google’s Language Model for Dialogue Applications (or LaMDA in short) are only further technological enhancements of a concept and model that is not novel in structure and principle but probably in design only.
Data Privacy considerations arising:
As artificial intelligence evolves, it amplifies the ability to use personal information in ways that can interfere with the privacy of an individual, such as facial recognition systems which are being deployed in airports, cities and various mobile apps and platforms worldwide. The use of personal data in AI for predictive policing, automated decisions and algorithmic decisions can indeed produce skewed algorithmic results or discriminatory results such as Amazon’s experiment with a hiring algorithm that simulated the company’s disproportionately male workforce and depicted a perceived bias against women[5]. Continued use of personal information about attributes such as skin colour, sexual identity, and national origin explicitly or implicitly for automated decision-making that could encroach on the privacy of the individual involved; highlights the further need to continue to update data privacy legislation to better address how personal data is used by developers whilst training their algorithms. In the alternative, lawmakers can ensure data privacy legislation such as the Kenyan Data Protection Act, 2019 is forward-thinking from a technology perspective and does not impede innovation.
Intellectual Property considerations arising:
Legislators worldwide will need to be alive of the salient patent, copyright and trademark policy questions arising, as AI development becomes increasingly sophisticated.
Regarding patentability, in 2022, the case of Thaler v. Vidal[6], arose from two patent applications filed in 2019 by Stephen Thaler, naming an AI system he called DABUS (for “Device for the Autonomous Bootstrapping of Unified Sentience”) as the “inventor.” The federal appeals court and the U.S. Patent and Trademark Office (USPTO) informed Thaler that the applications were incomplete because they did not list a human inventor. Whereas the Patent Act does not define an “individual,” the Federal Circuit cited a 2012 Supreme Court decision that held, in relation to text associated with a different statute, that an “individual” is normally understood to be a human being. The Federal Circuit concluded that in passing the Patent Act, “Congress has determined that only a natural person can be an inventor, so AI cannot be.” In Kenya, whereas the feasibility of naming an AI system as an inventor of a patent has not yet been tested by a court of law;- an “inventor” in Kenya means the person who devises the invention as defined in section 21 of the Industrial Property Act and includes the legal representative of the inventor. In this regard, the Kenyan position will likely mirror the U.S. position in terms of inventorship.
The best outcome could be declaring that the use of AI should not render unpatentable and otherwise patentable inventions, and attribute inventorship to the natural persons behind the AI. Declaring an AI system as the inventor of a patent would require an overhaul of the current framework around patentability.
Trademark infringement usually depends on whether there would be a likelihood of confusion caused by the defendant’s use of an identical or similar sign to a registered trademark. The aspect of brand recognition and the likelihood of confusion changes where AI is involved in situations such as AI suggesting products based on browser history and data analysis. This, therefore, suggests that the burden of proof and ingredients to be proven by plaintiffs in trademark infringement cases will require rethinking/review where a trademark infringement case is premised on an AI system/technology making decisions and purchasing products on behalf of a human. There is increased debate[7] about whether the doctrine of the average consumer is applicable to machines such as Alexa and other bots, as they slowly replace the natural “average” consumer in decision-making and purchases. This position is yet to be tested in Kenya, and in the same vein, the Trademarks Act, Cap 506 does not make provision for an AI system or algorithm to be registered as the proprietor of a mark in Kenya.
In Kenya, the Copyright Act, Cap 130 (the Copyright Act), does not yet recognize copyright ownership or authorship by an AI system. The Copyright Act clearly states that an author in relation to artistic work or a computer program which is computer generated, means the person by whom the arrangements necessary for the creation of the work were undertaken (a natural person). The Copyright Act goes ahead to state that an author in relation to a computer programme, means the person who exercised control over the making of the program (a natural person). In the United States and several other jurisdictions, copyright authorship can only be granted to works that were created by a human and that are sufficiently original, in addition to other prescribed requirements under current legislation.
Therefore, for patentable and copyrightable works to be attributed to an AI/machine system in Kenya and worldwide, current legislation will require review and amendment to stay in touch with these winds of change in technological developments.
[1] https://news.microsoft.com/source/features/ai/microsoft-approach-to-ai/
[2] https://web.archive.org/web/20170827035527/http://wsoe.org/google-collects-ai-based-services-across-the-company-into-google-ai/
[3] https://openai.com/blog/chatgpt
[4] https://blog.google/technology/ai/bard-google-ai-search-updates/
[5] https://www.reuters.com/article/us-amazon-com-jobs-automation-insight-idUSKCN1MK08G
[6] https://cafc.uscourts.gov/opinions-orders/21-2347.OPINION.8-5-2022_1988142.pdf
[7] Cosmetic Warriors and Lush v Amazon.co.uk and Amazon EU ([2014] EWHC 181 (Ch))
