June 4, 2023
-
Regulation
3 Obstacles to Regulating Generative AI
Harvard Business Review, 10/31/23. Governments are taking steps to regulate generative AI in order to mitigate risks. However, these regulations may face challenges in their effectiveness. Three obstacles need to be overcome for successful regulation: the complexity of AI systems, the rapid pace of technological advancement, and the lack of international consensus on AI governance. Instead of solely focusing on regulations, it may be more effective to regulate the development processes of generative AI and embed laws within software systems. This would ensure compliance and accountability throughout the AI development lifecycle. READ THE ARTICLE
-
Research
DeepMind touts AlphaFold’s new skills as protein-folding AI models face off
Stat News, 10/31/23. Google DeepMind and Isomorphic Labs have made a significant update to their deep learning model, AlphaFold. In 2020, AlphaFold revolutionized computational biology by predicting protein structures from their amino acid sequences. Now, the updated model can also predict the structure of proteins when combined with other molecules, such as small molecules and nucleic acids. This breakthrough has important implications for drug development, as it allows for the design of more effective drugs. Additionally, the new AlphaFold can predict protein structures even after they have been modified. This development is highly impressive and showcases the potential of deep learning in advancing scientific research. READ THE ARTICLE
-
IP
Court Refuses to Dismiss Artist’s Copyright Claim Against Stability AI
The Fashion Law, 10/31/23. In a recent AI-centric lawsuit, a federal judge in California has ruled in favor of Stability AI, DeviantArt, and Midjourney, dismissing most of the claims made against them by three artists. However, the judge allowed the artists to amend their complaint and clarified that one artist’s direct copyright infringement claim against Stability AI can proceed. The court also dismissed claims of vicarious copyright infringement, Digital Millennium Copyright Act violations, right of publicity, unfair competition, and breach of contract. This ruling provides insight into how courts may handle future cases involving generative AI platforms and copyright infringement. READ THE ARTICLE
-
Education
Students Outrunning Faculty in AI Use
Inside Higher Ed, 10/31/23. According to a new report, faculty members have been slower in adopting artificial intelligence (AI) tools compared to students. The report shows that nearly half of college students are already using AI tools, while only 22% of faculty members use them. The study also reveals that faculty members who are familiar with AI recognize its importance, with 75% of regular AI users believing that students need to know how to use generative AI in a professional setting. However, the report indicates that faculty members are taking their time with both adoption and policy implementation, with many still learning how to use AI tools. The report suggests that the differential between student and faculty use of AI will be something to monitor. READ THE ARTICLE
-
IP
EFF to Copyright Office: Copyright Is Indeed a Hammer, But Don’t Be Too Hasty to Nail Generative AI
Electronic Frontier Foundation, 10/31/23. In recent court rulings, important decisions have been made regarding copyright infringement and fair use. A federal appeals court ruled that technical standards incorporated into public law can be freely disseminated without liability for copyright infringement. This is a significant victory for open access to information. Additionally, a decision by the Supreme Court affirmed fair use precedents in a case involving commercial artists. These rulings support the rights and obligations of artists and creators in the digital age. It is encouraging to see the courts protecting freedom of expression and innovation. READ THE ARTICLE
-
Politics
AI Modi started as a joke, but it could win him votes
Rest of World, 10/30/23. The rise of AI voice-cloning tools in India has allowed meme creators to mix politics with entertainment and trolling, drawing in large audiences. These videos, featuring Indian Prime Minister Narendra Modi’s digitally rendered voice, help break down language barriers and make Modi’s campaign accessible to a wider population. With the 2024 Indian election expected to have 1 billion voters, AI is set to play a significant role in influencing political decisions. However, the use of synthetic media and deepfakes also raises concerns about internet regulation and privacy. READ THE ARTICLE
-
Marketing
This wild AI tool can turn any website into a better version of itself
Fast Company, 10/30/23. The article introduces Adept, a startup leveraging AI to revolutionize website interfaces and enhance user experiences, using Airbnb as a prime example. Adept’s approach involves using AI, including Large Language Models (LLMs), to read and interpret webpage content in real-time, creating personalized overlays for users. This innovation aims to make websites more interactive and user-focused. Additionally, the concept of “zooming in and out” in UI design is explored, offering users varying levels of information and functionality. Adept’s vision is to democratize and customize web interfaces, ultimately improving accessibility and usability for all users. READ THE ARTICLE