AI News

Experts suggest open-source large language models (LLMs) may surpass proprietary models in enterprise use, although real-world applications have been limited. Companies prioritize data privacy and control, leading to cautious adoption. Currently, 16 notable use cases exist, with expectations for growth as organizations explore open-source solutions.

Acadian Ventures has raised $30 million for its second fund, focused on technologies that enhance work. Backed by ServiceNow Ventures and others, the fund targets intelligent applications, work infrastructure, compliance solutions, and the global workforce, having already invested in 12 startups.

At Davos 2024, discussions around AI shifted from optimism to caution, focusing on risks like misinformation and job displacement. Key leaders stressed the need for trustworthiness and regulatory frameworks amid growing concerns over deepfake technology impacting democracy. Public sentiment remains mixed, reflecting a need for informed guidance as organizations explore AI's potential while ensuring it serves humanity positively.

On January 26, 2024, SAG-AFTRA addressed AI scandals involving deepfake images of Taylor Swift and a comedy special featuring George Carlin. They called for regulations against unauthorized images and supported bipartisan legislation to combat non-consensual deepfakes. The urgency for reform is escalating.

The AI revolution will enhance the role of data engineers by requiring new skills for effective implementation. Key transformations include efficient data pipeline building, reduced data mapping, upskilling BI analysts for dynamic reporting, and managing third-party AI services, leading to more strategic roles in data engineering.

The estate of comedian George Carlin is suing Dudesy over an AI-generated comedy special featuring his likeness, claiming publicity rights violations and copyright infringement. Critics call the content subpar, and Carlin’s daughter rejects the idea of his “resurrection” through AI. The lawsuit seeks damages and removal of the special, highlighting concerns about AI's impact on creative expression.

A cnvrg.io survey reveals slow enterprise adoption of AI, with only 10% deploying generative solutions. Key barriers include infrastructure issues and skill shortages, impacting sectors like education and automotive. While 46% cite infrastructure as a primary hurdle, 84% recognize the need for better skills.

Biometric data threats, including deepfakes and compromised identities, are surging on the dark web, targeting C-level executives vulnerable to phishing scams. Current security measures are insufficient. Badge Inc. aims to enhance biometric security with innovative, cryptography-based solutions, supporting a zero-trust architecture.

Generative AI is transforming web browsing through new features from privacy-focused browsers like Arc and Brave. Arc integrates the AI search engine Perplexity, while Brave upgrades its assistant Leo with Mistral's Mixtral LLM. These innovations aim to enhance user interaction and privacy, signaling a shift towards an AI-driven browsing experience.

AI startup Anthropic faced a week of scrutiny after a contractor mistakenly shared non-sensitive customer info with a third party. The incident, deemed human error, triggered FTC investigations into Anthropic's partnerships with Amazon and Google, raising concerns about market competition and data security.

AlphaCodium, a new open-source AI code generation tool from CodiumAI, improves on Google DeepMind’s AlphaCode by using an innovative "flow engineering" method, enhancing code integrity via Generative Adversarial Networks. It has outperformed existing models and increased GPT-4's accuracy significantly. Founded in 2022, CodiumAI raised $10.6 million to support its mission of producing bug-free code efficiently.

Skill Money Games launched an AI-driven platform for cash-based golf challenges, enhancing the indoor golf experience. Backed by $5 million in funding, the app offers personalized contests for varying skill levels. The company plans to expand partnerships and explore AI applications in other sports.

Microsoft has updated its AI assistant, Recall, to enhance security and privacy protections. While Recall is optional and can't be fully uninstalled, users can choose to enable it during the setup of a new Copilot+ PC, which must meet strict hardware requirements. Users have control over data retention and can exclude specific apps from being saved. Data within Recall is encrypted, and access requires biometric authentication. Microsoft is also conducting penetration tests with a third-party vendor to validate its security measures, indicating a commitment to addressing user concerns about data privacy.

OpenAI is launching a new subscription tier, ChatGPT Team, priced at $30 per user monthly. It offers advanced features like GPT-4 with a 32K context window and collaboration tools. Targeted at businesses, it builds on the success of ChatGPT Enterprise, focusing on enhancing productivity and collaboration.

Researchers from Adobe and UC Berkeley developed an AI tool to detect facial manipulation in Photoshop- edited images. It uses a CNN to identify changes with 99% accuracy, helping restore trust in digital media and enabling users to recognize alterations effectively.