50% of Motiva AI R&D now dedicated to Responsible AI
PALO ALTO, CA, July 20, 2023 /24-7PressRelease/ — Motiva AI announced major advancements for Motiva Generator™, a generative AI engine for enterprise marketing teams. Motiva Generator now combines cutting-edge mixed language models, including hybrid mixed models like Llama2, Vicuna, Falcon, and Orca, customized enterprise Large Language Models (LLM)s, and proprietary closed models like OpenAI. In addition, Motiva Generator™now automatically adapts to enterprise customers’ proprietary data without sacrificing privacy or risking data leaks outside enterprise control.
McKinsey Global Institute estimates that generative AI products like Motiva AI will generate $2.6 to $4.4 trillion of annual value for the global economy. Goldman Sachs predicts these technologies to contribute $7T to global GDP. And while global enterprise CIOs are bullish on AI, with 95% currently trying AI, only 14% expected to adopt AI enterprise-wide by 2025. Clear gaps remain between trusted AI and enterprise adoption of generative AI.
“It’s the Wild West out there regarding generative AI usage. Enterprises are rightly asking how they can take advantage of all the innovation without sacrificing data controls, privacy, and security,” said David Gutelius, Motiva AI CEO. “Our DNA combines Enterprise customer engagement and national security, and Motiva Generator™reflects that experience in helping enterprises move fast and build protected, competitive advantage.”
Motiva Generator™now includes customizable model training and Motiva’s privacy-preserving, real-time query augmentation for model-shaping at run-time. These features ensure that enterprise-owned data, branding, and other protected assets are never used by others. Coupled with Motiva’s proven audience-based Reinforcement Learning from Human Feedback (RLHF), Motiva Generator™offers enterprise teams a full solution for constantly improving model outputs from trusted AI – whether for marketing, sales, customer support, or other dynamic use cases. Motiva Generator™roadmap adds adversarial model selection, automatically picking the right mixture of generative models for optimal performance.
“Motiva released its enterprise generative AI product in 2021,” said Gutelius. “We’re excited to bring Motiva Generator™and our latest work in RLHF to customer marketing teams.”
In addition to modeling advances for enterprises, Motiva announced a Responsible AI-first approach to data product development, with 50% of AI R&D now dedicated to bias, safety, and ethics in creating new AI capabilities. This includes not only compute costs but also engineering and research time commitment to Responsible AI, as well as a commitment to open-sourcing new Responsible AI products and tools as a contribution to the larger global community
“In the race to create new LLMs for generative AI-based products, ethics is a second-class citizen,” said Gutelius. “Instead of an afterthought, Motiva is making Responsible AI as important as anything else we do with AI. Taking a laissez-faire approach to safety, bias, and ethical development is not an option, and we can’t wait for politicians to regulate industry. The stakes for the world are too high.”
Motiva Generator™ is accepting beta signups now. For more information on Generator, visit https://www.motiva.ai/generator-beta/.
Motiva AI is a leader in Responsible AI for enterprise marketing. Motiva serves some of the most iconic brands in the world, including Blackrock, Agilent, PTC, and others, with AI-driven productivity tools that improve digital engagement and intelligence.
—
For the original version of this press release, please visit 24-7PressRelease.com here
