AI Rundown by Lightscape Partners – 12/25/23

Merry Christmas, happy holidays, and good morning! I hope everyone is enjoying the holiday. Like the three kings, who could not be stopped on their journey to Baby Jesus, nothing will stop us from delivering last week’s AI news to you! However, pending unforeseen ground-breaking news in artificial intelligence, the AI Rundown will resume January 8th.

This week's highlights include encouraging words from Bill Gates regarding the future of AI in 2024, Apple publishing breakthrough technology that drastically increases LLM efficiency on cellular devices, and a prominent dataset used in training models is taken down after discovering it contained illegal abuse images.

Top AI Developments of the Week

Bill Gates’ Optimistic Predictions for AI in 2024. Link.

In Bill Gates' article "The Year Ahead 2024," AI is discussed as a powerful tool for enhancing global equity and advancing innovation. Gates foresees significant adoption of AI in developed countries within 18-24 months and in African countries within three years.

• He emphasizes AI's potential to bridge the innovation lag between richer and poorer regions, underlining its transformative impact on global health and development.

• Gates also predicts AI will be a catalyst in the democratization of education, tailoring learning to each student’s “style”, mentioning tools like Khanmigo.

Apple Publishes Breakthrough Method Enabling LLMs to Run Efficiently on iPhones. Link.

Apple's researchers have published a new paper detailing a breakthrough method that enables large language models (LLMs) to run efficiently on devices with limited memory. This innovation is a significant step toward bringing powerful AI capabilities to handheld devices.

Amidst speculations about Apple's AI strategies, this development suggests that a range of new AI-powered features could be introduced in future iPhone models. These features are expected to enhance user experience and device functionality, leveraging the power of AI.

Apple's latest research in AI underscores its ambition to stay competitive and catch up with other tech giants in the rapidly evolving field of artificial intelligence. This move positions Apple as a key player in integrating AI with consumer technology.

LAION-5B Dataset Withdrawn After Researches Discovered the Dataset Contains Illegal Abuse Images. Link.

The LAION-5B, a vast dataset widely used for training AI models, has been temporarily pulled offline amidst concerns over its stability and potential for misuse.

Conducted by Stanford University’s Internet Observatory, the report claims that LAION-5B, a dataset containing lists of links to images throughout the internet, contain a “significant number of illegal abuse images.”

• Maintained by the non-profit LAION (Large-scale Artificial Intelligence Open Network), the dataset contains lists of links to images, not actual images themselves.

• After receiving the report, LAION immediately suspended the dataset.

• The organization behind LAION-5B is evaluating measures to prevent abuse and ensure the dataset's responsible use.

This move highlights the ongoing challenges in the AI field related to dataset governance and the ethical training of AI systems. The custodians of the dataset are working on protocols and safeguards to address these concerns before making it available again to the AI research community.

Venture

We are excited to be a part of the third largest funding round of October 2023, with Shield AI raising $200M in their Series F. We believe in the future of automated pilots and believe in the team at Shield AI to bring the best technology to the space.

The Biggest Raises in AI Startups for 2023. Link.

1. OpenAI, $10B

2. Anthropic, $4B

3. Anthropic (again), $2B

4. Inflection AI, $1.3B

5. Metropolis $1.1B

AI Industry News

Google Starts Replacing Human Staff with AI. Link.

According to a report by The Information, Google is looking to reassign or release sales employees whose positions were automated away.

• The report does not specify now many jobs were being replaced.

• Google begins a new ad initiative of “AI-powered ads”, with automatically created assets that scrape content from the web.

• The Performance Max feature can change in real-time based on click-through rates, maximizing visibility.

• This feature is extremely labor intensive for a human work, and Google once had 15,000 employees dedicated to this task.

Google’s move highlights the fear of many American’s regarding AI and some researchers think this adoption will continue to accelerate into the new year.

Leadership Restructuring at OpenAI. Link.

Organizational Update:

OpenAI has announced a reshuffle in its leadership, placing new board members in positions of oversight over Sam Altman, the organization's current CEO.

• The move marks a significant shift in the governance structure of OpenAI, aligning with its broadening influence in the AI sector.

• This change is designed to bolster strategic decision-making and enhance oversight as OpenAI scales up its operations.

The restructuring is expected to influence OpenAI's future strategy and its ongoing projects, ensuring robust leadership as the company navigates the rapidly evolving AI landscape.

With the new board members at the helm, OpenAI is poised to continue its pioneering work in AI research and development with renewed governance.

Regulation / Governance

New Jersey Announces AI Hub to Advance Innovation in Artificial Intelligence. Link.

New Jersey, in partnership with Princeton University, is establishing an AI hub aimed at advancing innovation in artificial intelligence. This initiative, leveraging state funding and private sector partnerships, is set to catalyze AI technological development, economic growth, and job creation in the region.

• Between Princeton University and the State of New Jersey.

• Based in New Jersey, utilizing state and university resources.

• Expected to drive regional economic development and job creation.

• Emphasis on ethical AI development and broad sectoral applications.

• Involvement of local educational institutions for workforce development.

• Aims to establish New Jersey and Princeton University as leaders in AI.

The establishment of the AI hub in New Jersey marks a significant step towards integrating advanced technology into the region's economic and educational fabric. It is poised to create new job opportunities, foster innovation across various sectors, and place New Jersey at the forefront of AI research and application.

UK Supreme Court to Rule on AI's Role in Patents

The UK Supreme Court is deliberating a significant case to determine if an AI system can be recognized as an inventor in patent applications.

• Dr. Stephen Thaler's AI system, DABUS, is the focus of this case, challenging traditional intellectual property concepts.

The court's decision will set a precedent for how AI-driven innovations are handled legally, reshaping the landscape of intellectual property rights in the era of artificial intelligence.

Ethics in AI

Anthropic Announces Legal Aid to Users Affected by Copyright Lawsuits. Link.

Anthropic announced in updated terms and services that the company will not only defend users of the Claud LLM from copyright infringement, but will also cover the costs of settlement, setting a precedent for user protections in generative AI.

• This follows similar patterns set by Microsoft, Google, and others.

• Protections do not cover a user who “knows or reasonably should know” they are infringing on copyright.

AI Voice Clone of Jailed Pakistani Prime Minister Raises Concerns. Link.

A recent development in AI technology has led to the creation of a convincing voice clone of Imran Khan, Pakistan's former Prime Minister, using ElevenLabs' AI software. This advancement has sparked discussions about the ethical implications and potential misuse of AI in impersonating public figures.

• ElevenLabs' AI technology was used to replicate Imran Khan's voice.

• The voice clone is reportedly highly convincing, leading to potential concerns about its misuse.

• This technology showcases the advanced capabilities of AI in voice synthesis and cloning.

The creation of Imran Khan's voice clone underscores the growing capabilities of AI in mimicking human voices, raising questions about the ethical use of such technology, especially in the context of public figures. It highlights the need for regulations and guidelines to prevent the misuse of AI in spreading misinformation or impersonating individuals.

Thank you for reading the AI Rundown by Lightscape Partners. Stay tuned for the latest updates in the world of Artificial Intelligence!