Today, businesses of all sizes and industries want to know how to use Generative AI and what is the safest way to do so. We at DaveAI have been helping brands adapt to the new waves of technology that keep coming in by analyzing their requirements, suggesting strategies and developing them.
We have recently launched our new Generative AI pipeline: DaveAI’s New GenAI Pipeline
With so much going on in this field, what are the “must knows” of generative AI for brands and what contributions does DaveAI make in this domain ? I asked our Co-founder & CEO, Sriram P H in an interview.
Excerpts Of The Interview:
1. Could you provide an overview of our company’s current research initiatives in the field of generative AI?
Sriram P H: At DaveAI, we invest heavily in R&D and have 3 research initiatives at present.
a. With Academia:
DaveAI partners with academic institutions like Plaksha University, IIIT Hyderabad and KTH Stockholm to exchange knowledge and ideas. Researchers from academia bring fresh perspectives and the latest research findings to the table. DaveAI helps provide real-world data and practical applications for research. Recently, students at Plaksha University explored a Metaverse use case by creating a platform that helps understand the role of Metaverse in education.
Working with academia helps address ethical concerns and promote responsible AI development. DaveAI’s research initiatives with academia aim to foster innovation, broaden knowledge, and drive the responsible development of AI technologies. These partnerships serve as a bridge between the academic and industrial realms, benefiting both sectors and society at large.
b. Technology Ecosystem:
DaveAI is actively engaged in developing a robust technology ecosystem for AI research and development. This involves creating a supportive infrastructure that encompasses software, hardware, and frameworks. The ecosystem includes designing specialized hardware, developing software libraries, and optimizing AI algorithms to work efficiently together. This ensures that AI models and solutions can run smoothly and deliver high performance.
DaveAI has formed a strategic partnership with Intel, a leading technology company known for its innovations in hardware and processors. This partnership is centered around exploring the specific hardware requirements for AI applications. Intel’s expertise in processor architecture and hardware acceleration is invaluable for optimizing AI workloads. Together, Intel and DaveAI are conducting research to determine the most suitable hardware configurations. DaveAI’s collaboration with Intel to explore hardware requirements goes beyond pure research.
It also has a practical application in delivering AI solutions for brands. By understanding the hardware requirements, DaveAI offers brands customized AI solutions that are not only powerful but also cost-effective and energy-efficient. These solutions include AI models for tasks like natural language processing, computer vision, and recommendation systems, all optimized to run on the hardware configurations identified through the partnership with Intel. Brands benefit from these tailor-made AI solutions that can enhance their products and services, improve customer experiences, and drive business growth.
c. In house research:
DaveAI is committed to enhancing the existing AI products and developing new ones. In-house research plays a pivotal role in achieving this goal. The research team at DaveAI focuses on identifying areas where product improvements are needed. This involves refining algorithms, optimizing performance, improving user interfaces, or adding new features. Continuous innovation is essential in the fast-evolving field of AI. The in-house research team works to stay ahead of the competition and ensure that DaveAI’s products remain cutting-edge and competitive in the market.
In-house research also encompasses rigorous testing and quality assurance. This research helps identify and address potential issues, such as bias in AI models, security vulnerabilities, or scalability challenges. It’s crucial for delivering products that meet high standards of quality and reliability. In-house research allows us to adapt to these changes swiftly.
DaveAI’s dedication to responsible AI extends to our commitment to deliver high-quality experiences while minimizing environmental impact. In an era where the environmental consequences of technology are a growing concern, we are actively researching and innovating to find sustainable solutions. Our focus is on optimizing AI algorithms to operate efficiently on low GPU resources and reducing our reliance on massive data centers. By doing so, we aim to reduce the carbon footprint associated with AI technologies, making them more eco-friendly and aligned with global sustainability goals.
2. Can you share some insights achieved by our research team in generative AI?
Sriram P H : Speaking of our latest Gen AI pipeline, To incorporate generative AI in our avatars, the first step was to research about how these avatars converse with customers of brands and vice versa. There were two aspects to this. While developing this solution we tried to answer 2 important queries:
a. How does a user create content with DaveAI avatars?
The user asks a question, for which the NLP system forms a response, feeds it to the avatar and then the avatar animates the response with hand movements, gestures, facial expressions to explain it to the user.
b. How can DaveAI help enterprises adopt the power of LLMs without compromising on data security?
Brands can adopt LLMs and still ensure that their data remains safe and secure. To do this, brands need to incorporate DaveAI’s Generative AI ecosystem as the middleware. DaveAI’s named entity recognition system as the middleware filters data to ensure sensitive information is not used and also controls the responses that go out from the brand by creating responses that are in tune with ethical considerations and also in tune with the brand messaging.
The middleware layer enables brands to fine-tune the tone and style of messages generated by the LLM to align with their specific brand voice. Brands can dynamically adjust the middleware settings to adapt to changing marketing campaigns or customer trends. Using the middleware layer, brands can maintain a consistent tone and messaging style across various channels and platforms, reinforcing their brand identity. The middleware also includes checks for compliance and ethical considerations, ensuring that generated content adheres to legal and ethical guidelines.
3. What role does ethical AI play in our AI solutions and what measures are in place to ensure responsible AI development?
Sriram P H : The notion that AI models have the potential to become the arbiters of truth raises critical questions about the concentration of power and the need for vigilant oversight. In an era where AI systems play increasingly influential roles in decision-making, from content curation to data analysis and beyond, there is a growing concern that the algorithms governing these models could inadvertently or intentionally perpetuate bias or misinformation. As AI’s influence expands across various aspects of society, it becomes imperative to address the ethical, regulatory, and accountability challenges associated with these technologies. This underscores the importance of striking a balance between technological advancement and preserving transparency to ensure that AI remains a tool that serves the greater good rather than concentrating power in the hands of a few.
In an age where AI-generated content blurs the lines between reality and fabrication, the ability to discern what is genuine from what is AI-generated holds immense significance. At DaveAI, our commitment to ethical AI development is exemplified through our creation of photo realistic avatars using technology. Unlike deepfake, our approach prioritizes transparency, integrity, and responsible AI usage. By crafting lifelike avatars without resorting to manipulative techniques, we aim to foster trust and reliability in AI applications. This dedication to ethical AI not only safeguards against potential misuse but also ensures that our technology serves as a constructive force, enhancing the human experience rather than undermining it.
Taking inspiration from Meta’s open source approach to LLMs, we plan to adopt the same strategy for our platform as well and allow developers to build on top of it. This means opening doors for developers to harness the power of our technology. By fostering an open ecosystem, we aim to encourage innovation, collaboration, and the co-creation of new applications and solutions. This approach not only democratizes AI but also promotes transparency and accountability, ensuring that AI development remains a collective effort for the benefit of all.
DaveAI’s founding philosophy centers around the mission to democratize AI. This means making artificial intelligence accessible, understandable, and usable for everyone, regardless of their technical expertise or resources. This philosophy reflects a commitment to removing barriers that might otherwise limit the benefits of AI to a select few, with the ultimate goal of empowering individuals and organizations to harness the potential of AI for their specific needs and purposes.
4. In what ways does our company ensure that the research findings are translated into practical solutions for our customers and clients?
Sriram P H: At DaveAI, our approach to problem-solving is rooted in a systematic and customer-centric research methodology. We begin by closely examining the challenges our customers face, delving into their unique needs and pain points. This initial research serves as the foundation for our problem-solving process. Once we have identified the specific challenges, we embark on a comprehensive analysis to determine the most suitable solutions for each scenario. We scrutinize existing offerings in the market to evaluate whether adequate solutions already exist or if there is room for innovation.
Our commitment does not stop at identifying problems and solutions; we take a proactive stance in developing optimal, customized products. This involves a meticulous assessment of product efficiency and its potential for expansion across various business use cases. We continuously monitor and assess the business outcomes of our research and development initiatives, ensuring that our efforts not only address immediate concerns but also contribute to long-term success and growth.
To use customer data we use meta level learning. The fundamental idea behind meta-learning is to train a model on a variety of tasks in such a way that it becomes skilled at learning new, unseen tasks with minimal data or training. This involves learning a set of high-level patterns or strategies that can be applied across different tasks.
5. What future trends or innovations do you foresee in the field of generative AI?
Sriram P H: The future prospects of generative AI point towards large organizations and billion-dollar companies increasingly investing in and creating their own AI models. This trend is already visible today, with tech giants like Google, Facebook (now Meta), OpenAI, and others developing proprietary AI models. As AI models continue to advance in capability, there’s ongoing speculation and discussion about the potential for AI to achieve super intelligence. Having exclusive access to advanced AI models can provide a significant competitive edge, as it can lead to innovation in products and services and create barriers to entry for competitors. For companies with diverse operations, owning AI models can facilitate integration across various business units, streamlining processes and decision-making.
In the ever-evolving landscape of artificial intelligence, it is increasingly evident that the direction and impact of AI systems is profoundly influenced by human decisions and oversight. As AI becomes more pervasive, the role of humans in guiding, supervising, and setting the intent for AI adoption becomes paramount. Humans are responsible for defining the objectives, values, and boundaries within which AI operates. It is humans who must ensure that AI systems align with societal norms, adhere to ethical principles, and respect individual rights.
