CALI, COLOMBIA - APRIL 17: A Colombian urban shaman plays the cards in orther to predict the future ... [+]
AI is real. But despite the fact that Artificial Intelligence (AI) and Machine Learning (ML) are quite definitely with us, a certain level of AI agnosticism is still required. More specifically, we need to adopt an agnostic stance with regard to the components that come together to build our AI smartness. We are of course talking about the arrival of new and updated Large Language Models (LLMs), which underpin AI knowledge in their various shapes, sizes and guises.
As much as organizations now have an opportunity to deploy enterprise AI applications that have the potential to change business operations, it is critical for any company evaluating AI solutions to understand how they work with the various LLMs available and which will offer the biggest advantages.
As such, AI agnosticism will be vital.
What is AI agnosticism?
Not exactly a formally practice, industry-standard approach or defined methodology as such, AI agnosticism is like any form of IT agnosticism in that it advocates the adoption of generalized interoperable technologies, process, practices, tools and components. Applicable to both hardware and software, IT agnosticism and AI agnosticism in particular means keeping an open mind (figurately and literally) in relation to what data knowledge resources we use to build AI models for any number of business use cases.
BlueFlame AI, providers of a generative AI platform for alternative investment managers, believe now is the time for firms to embrace LLM-agnosticism and to look for AI solutions that are Large Language Model (LLM) and AI provider agnostic. BlueFlame deploys its LLM agnostic platform to help firms select the best LLM for specific tasks that can reduce risk and optimize performance and efficiency.
“The facts are straightforward - we can say that optimization, customization and resilience are three critical benefits LLM-agnostic AI solutions can deliver, with optimization being key as it allows users to select the right model for each specific task based on model strength and weaknesses,” asserts James Tedman, head of Europe for BlueFlame AI. “Businesses using LLM-agnostic AI solutions will stand to have the greatest impact because they will reduce model biases and decrease reliance on any one single LLM. If there is an LLM outage, businesses will want diversified LLMs to pivot to an alternative without significant disruptions. LLM-agnostic approaches ensure your business doesn’t have to deal with service discontinuation or performance issues.”
A volatile time in AI
An LLM agnostic approach requires that software developers understand each LLM’s capability, limitations and complexities. The volatility of some of the AI events (job moves, fast-moving platforms, acquisitions and so on) in 2023 events arguably showed us how risky reliance on one LLM could be. It is suggested that an agnostic approach (or at least a more agnostic approach) can ensure reduced disruption from product or pricing changes, service disruptions, or more. It’s this type of adaptability that’s critical in today’s evolving AI landscape.
“Various models might require different strategies for data management or processing. Developers will need robust Application Programming Interfaces (APIs) and middleware solutions due to the increased complexity of multiple LLM integrations,” said Tedman. “Since each has its own learning behaviors, it’s important to ensure LLMs are providing consistent and reliable results.”
Enterprise AI guardrails
Staying secure and compliant will be another key priority for any business building enterprise AI applications. Regulators are already gearing up to review the space, especially following the EU’s AI Act. So then, what best practices can businesses follow to make sure they’re compliant? Tedman explains that regularly updating security protocols to protect against sensitive information processed by various LLMs will support data security.
“Ensuring that you have commercial agreements in place to prevent data being used for model training and adherence to privacy laws like GDPR or CCPA will support privacy requirements. It will also be critical to implement strict access controls and authentication mechanisms to prevent unauthorized access to AI systems or sensitive data,” he noted.
The BlueFlame team have reiterated that regularly auditing AI systems for security vulnerabilities and monitoring for malicious activities will keep organizations vigilant. Companies should also ensure that their AI solutions comply with industry-specific standards and regulations, particularly in sectors like financial services, healthcare and legal.
“It’s vital that businesses maintain transparency in how their AI systems use and process data, informing stakeholders about the AI models in use and their data handling practices. Businesses that can leverage LLM agnostic AI solutions to optimize performance and mitigate risk stand to have the greatest impact in the upcoming AI revolution,” concluded Tedman.
Greek for knowledge: gnōsis
It is perhaps no coincidence that - as Gavin Wright on TechTarget explains here - the word agnostic comes from the Greek a-, meaning without and gnōsis, meaning knowledge. “In IT, that translates to the ability of something to function without ‘knowing’ or requiring anything from the underlying details of the system it is working within.”
We may evolve the term AI agnosticism to become AI openmindedness, AI rationalism or perhaps even faithless AI in future, who knows. What we can say is that believing in AI is important, but believing in non-secular AI that is open to all information sources might well be more important.
Let us AI.
