AI News

UG Labs, based in Tel Aviv, has raised $7 million to integrate conversational AI into children’s games, enhancing voice interactivity. Their technology utilizes ASR and LLM to foster dynamic, safe dialogues, promoting engagement and personalized experiences for young users.

Elon Musk will open source xAI's Grok chatbot to encourage technological sharing and boost adoption, echoing Tesla's past actions. This move could challenge OpenAI, with Musk's lawsuit alleging it prioritizes profit over its original mission. OpenAI claims Musk attempted a merger with Tesla.

Anthropic recently launched Claude 3.0, advancing AI capabilities with enhanced safety and multimodal functions. Despite claims of nearing artificial general intelligence, experts see limitations. The competitive landscape is evolving, raising ethical concerns as AI technology continues to grow.

Zelda Williams criticized the unethical use of celebrities' digital likenesses without consent, highlighting concerns amid companies like Soul Machines creating AI chatbots like "Digital Marilyn." While marketed for engagement, the practice raises ethical issues, particularly around consent and exploitation.

AI's rapid integration raises concerns about misinformation, bias, privacy issues, and job displacement. Naming conventions for AI, often anthropomorphic and gendered, can reinforce stereotypes and impact user perception. Clear, non-anthropomorphic naming is essential for effective AI branding and communication.

Lenovo's "AI for All" initiative, showcased at Mobile World Congress 2024, emphasizes edge computing to empower first responders, reduce cloud energy use, and enhance operational efficiency. Collaborations with Telefonica focus on real-time AI applications in urban settings and retail, driving innovations in safety and customer experiences.

Microsoft's Copilot tool has restricted prompts to block the generation of violent and inappropriate images following concerns raised by engineer Shane Jones. Users can no longer enter terms like "pro choice" or "pro life," facing warnings and potential suspension for violations. Microsoft is enhancing safety filters to address these issues.

Sam Altman has rejoined OpenAI's board after a brief ousting as CEO. Following an investigation into his dismissal, three new members were added to the board. The inquiry clarified that his removal stemmed from a breakdown of trust, not safety or financial concerns.

Sam Altman, previously ousted as CEO of OpenAI amid leadership concerns, will rejoin the board after an investigation found no misconduct. The board has appointed three women members to enhance diversity. Altman’s return marks a pivotal moment as OpenAI faces scrutiny over its governance and AI ethics.

Insilico Medicine has raised over $400 million for AI-driven drug development. Their new paper in *Nature Biotechnology* details INS018_055, the first AI-generated drug in Phase II trials for idiopathic pulmonary fibrosis. Utilizing generative AI, its development was accelerated, demonstrating potential in drug discovery.

A Georgia Tech study reveals that large language models (LLMs) exhibit bias favoring Western culture, even when trained on Arabic data, raising concerns about cultural fairness in AI. The new benchmark, CAMeL, assesses these biases and aims to improve LLMs' cultural awareness and accuracy.

Adobe has launched its Firefly generative AI in a beta version of the Express app for Android and iOS, offering mobile users advanced creation and editing features. The app includes tools for image and video editing, collaboration, and integrates with Creative Cloud. Premium features are currently free, with a future subscription fee of $10/month.

Earlier this year, KoBold Metals made a groundbreaking discovery, unveiling what could be one of the largest high-grade copper deposits ever found. This deposit has the potential to yield hundreds of thousands of metric tons annually, according to the company’s CEO.

Grok, a chatbot founded by Elon Musk, exhibited significant vulnerabilities in safety tests, ranking lowest among major AI models for responding to harmful prompts. It easily bypassed safeguards against illegal activities, prompting calls for rigorous red teaming to enhance AI security and effectiveness.

Apple's upcoming iOS 18 feature, Visual Intelligence, enables users to quickly learn about items by pointing their iPhone's Camera Control button at them. It uses on-device intelligence for instant data retrieval without storing images, and will support third-party searches. Launch is expected later this year.