This Week in AI: MIT Database and Google’s Assistant Debut as Regulators Eye Apps

MIT Database, Google’s Assistant Debut as Regulators Eye Apps

From MIT’s artificial intelligence risk database to Google’s enhanced AI assistants, the tech world is abuzz with AI developments promising progress and potential pitfalls.

As companies like Anthropic and Google roll out efficiency-boosting features, regulatory bodies from the Federal Communications Commission to the Consumer Financial Protection Bureau are stepping up efforts to ensure AI’s responsible implementation across industries.

MIT Unveils AI Risk Database

MIT’s FutureTech Group unveiled an “AI Risk Repository,” compiling 777 potential AI pitfalls from 43 taxonomies. The database is poised to reshape corporate AI strategies, potentially slowing adoption while enhancing safety measures in an era of rapid AI proliferation. The repository’s launch coincides with global efforts to regulate AI, including the European Union’s AI Act and emerging U.S. state legislation. Experts predict it will become a cornerstone for policymakers and businesses navigating the complex terrain of AI implementation and governance.

Anthropic Launches Prompt Caching to Boost AI Efficiency

Anthropic unveiled “Prompt Caching,” a cost-saving and performance-enhancing feature for its Claude 3.5 Sonnet and Claude 3 Haiku language models. The innovation allows businesses to store and reuse contextual information in prompts without recurring costs or increased latency. As the AI industry heats up with rivals like OpenAI, Google and Microsoft vying for dominance, Anthropic’s latest offering aims to make its models more efficient and user-friendly, potentially giving it an edge in the competitive landscape of large language models.

Google Boosts AI Overviews: More Links, Wider Reach

Google has upgraded its AI Overviews feature, making links more visible and expanding to six new countries. The update aims to help users find content from publishers, businesses and creators more easily. AI Overviews now shows links prominently on the right side for desktop users and as site icons on mobile. Google is also testing in-text links within the AI-generated summaries. Hema Budaraju, a senior director at Google, said the feature is leading users to explore a broader range of websites for complex questions, with visitors spending more time on the sites they click through to from search results.

CFPB Eyes Bank Chatbots: Progress Meets Regulatory Scrutiny

The evolution of AI-powered chatbots in banking faces a new challenge as CFPB prepares to regulate these digital assistants. While banks have been using chatbots to enhance customer service and streamline operations, the CFPB’s impending crackdown aims to address concerns about their effectiveness and transparency. The regulatory scrutiny comes as financial institutions continue to refine their chatbot capabilities, with some, like Bank of America’s Erica, showing user growth and expanded functionalities. The move highlights the balance between technological innovation and consumer protection in the financial sector, potentially reshaping how banks deploy and develop their AI-driven customer service tools.

FCC Proposes AI Disclosure Rules for Robocalls

The FCC unveiled a proposal requiring businesses to disclose their use of AI in automated calls and texts, potentially reshaping customer outreach strategies. The move, aimed at enhancing transparency and combating fraud, could prompt companies to revamp their communication processes and invest in new technologies. Trevor Francis, CEO of 46 Labs, told PYMNTS businesses may need to adjust their scripts, reconsider automation use and explore alternative customer communication methods to comply with the proposed rules. As the FCC seeks public input on this initiative, companies face challenges and opportunities in adapting their customer interaction approaches to meet evolving regulatory standards.

Google Launches Gemini Live AI Assistant for Mobile Devices

Google introduced Gemini Live, a new AI-powered mobile assistant that promises to revolutionize smartphone interactions. Launched for Android users subscribed to Gemini Advanced, with iOS and multi-language support coming soon, the tool offers a conversational experience. Sissie Hsaio, a vice president at Google, highlighted Gemini Live’s ability to handle interruptions, pause-and-resume conversations, and operate in the background or with locked screens. The company plans to expand Gemini’s capabilities by integrating it with various Google apps and services, including Keep, Tasks, Calendar and YouTube Music, enhancing its utility across the Google ecosystem.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.