AI News

Paris-based Bioptimus has emerged with $35 million in seed funding to develop a universal AI foundation model for biology, leveraging generative AI across various biological scales. Backed by former Google DeepMind and Owkin scientists, Bioptimus aims for comprehensive modeling by accessing diverse patient data.

The rise of generative AI tools has led to increased fraud, including a $25 million scam in Hong Kong. Hummingbird, a San Francisco startup, is developing AI-driven automation software to aid financial crime investigators by streamlining compliance tasks and improving efficiency.

ElevenLabs, an AI startup specializing in voice cloning, is launching a text-to-sound model that generates sound effects from user descriptions, enhancing digital content. Currently, it's only available for early access sign-up. The market for AI audio tools is projected to grow significantly, reaching nearly $5 billion by 2032.

OpenAI's new text-to-video model, Sora, generates videos up to one minute long and exemplifies the company's mission toward developing AI that benefits humanity. While Sora dazzles as a creative tool, it represents a broader approach aimed at advancing artificial general intelligence (AGI).

Sam Altman’s recent employment changes and OpenAI's Q* model highlight the evolving landscape of artificial general intelligence (AGI). Deep learning, while crucial, has limitations due to its reliance on vast data and fixed predictions. To advance AGI, methodologies like Decision-Making under Deep Uncertainty may offer more robust, adaptable frameworks.

AI advancements require data professionals to adapt and enhance their roles in organizations. Despite concerns from 77% of business leaders, skilled data stewards are vital for effective data management. AI empowers CDOs, data architects, engineers, scientists, analysts, and developers, improving data quality and operational efficiency.

A coalition of 20 tech companies, including OpenAI and Google, has formed an agreement to combat AI deepfakes in the 2024 elections across 40 countries. The "Tech Accord" aims to mitigate risks and promote transparency, though concerns about its vague enforcement linger. Critics worry voluntary commitments may not be effective without stricter regulations.

The FTC warned companies against secretly changing Terms of Service to misuse user data for AI training, as such practices may be deemed unfair. Recent incidents highlight user ignorance about changes, with many firms facing scrutiny over consent protocols for data usage in AI.

OpenAI has launched the **OpenAI Forum**, an invitation-only community for domain experts and students to collaborate on AI. Membership is free but selective, requiring an hour of participation quarterly. Members engage in discussions, evaluations, and events to influence AI development and promote AGI advancement.

Join the AI Impact Tour in NYC on February 29, tailored for enterprise decision-makers. Engage with experts on the benefits and risks of generative AI, explore responsible implementation, and hear insights from leaders at NewYork-Presbyterian, Citi Bank, and Microsoft. Apply now!

*Frontiers in Cell and Developmental Biology* faced criticism for publishing an article with inaccurate AI-generated imagery of testicles and sperm cells. The journal retracted the article after community concern, raising questions about AI's impact on scientific accuracy and peer review practices.

Gong's AI-driven revenue intelligence software significantly enhances enterprise sales teams' productivity and win rates. Analysis of over 1 million sales opportunities showed that tools like Smart Trackers and Ask Anything increase win rates by up to 35%. Gong focuses on accuracy by training models on real sales data, aiding reps in automating tasks and improving customer interactions.

Meta’s chief AI scientist, Yann LeCun, continues to advocate for non-generative AI models, as evidenced by the recent announcement of the latest iteration of the Joint-Embedding Predictive Architecture (JEPA). This innovative approach prioritizes predictive learning over generative methods, offering a fresh perspective on how machines can replicate human-like learning processes.

Researchers at the University of Zurich developed an algorithm that enables an AI drone to outperform human pilots in racing, marking a significant advancement in autonomous drone technology. This breakthrough could improve efficiency in rescues and deliveries, with future enhancements aimed at reducing computational intensity and utilizing onboard cameras.