Ad hoc deployments of AI without the right measures and controls in place can be really harmful, so I'm glad to see this kind of legislation move forward.
Over the years, at scale, I saw firsthand how important these controls are when suicide notes appeared in chats. If you don't have the right monitoring capabilities in place or have designed your applications to accommodate the limitations of the models you're using, the result could be embarrassing for your brand at best, catastrophic for someone's family and friends at worst.
As developers and innovators, it's our responsibility to ensure we build and deploy applications that ensure this technology is not misused to cause harm.
Sad to see the opposition and surprised to see the supporters. Regardless, responsible use of technology (AI or otherwise) is a core ethical responsibility of building software.
#generativeai#ai#sb1047#legislation
"Inclusive" is one of four NLX company values. We believe that people may not always remember what you said, but they will always remember how you made them feel. Be approachable and make each other feel welcome. We expect diverse experiences and value individual input. Check out our other values here: https://nlx.ai/about
Deploying and supporting content and technical features across multiple languages and digital properties can be a significant challenge for global enterprise organizations.
But not when they work with NLX. Integration with one website property in one language is as fast and seamless as integrating with 20 properties in 20 languages. In fact, NLX supports over 70 languages and locales.
The AI community is rightfully abuzz with NVIDIA earnings results. However, regardless of the results from this remarkable company today, we believe the AI industry is headed for a slowdown. Why?
There’s a fundamental misunderstanding of AI technology among investors and AI customers, which has led to the gross misalignment of expectations of what it would do for businesses. Consider that reports estimate that as little as 10% of AI projects go into production. The highest estimate we’ve seen is 30%.
AI is based on statistics, and the effectiveness of AI applications on any business process depends on the content, data quality, and other inputs from the business. Despite what we see in the movies, AI does not have human-like cognition that can solve any complex business problem we can throw at it.
Additionally, operationalizing any new technology in a large organization is complicated and resource-intensive.
To align expectations and reality, investors and more people inside enterprise organizations need to better understand what AI is good at and what it’s not. When that happens, companies can be more strategic and precise about what AI applications to invest in, with proper expectations of return.
Consider that using large language models (LLMs) can be over 10X more expensive than its precursor, natural language processing (NLP) technology. So, companies deploying applications using LLMs without clearly understanding their value faced big price tags without concrete results.
At NLX, we've seen more customers concerned about cost early on, which is a good thing. However, our approach combines a creator-level understanding of the technology with a results-driven practical mindset. That means we only use LLM processing when necessary and when it can deliver a predetermined business result that justifies the cost.
Many other solutions in the market made a hard pivot to LLM-only assistants, which is both expensive and impractical because most business processes require 100% logic certainty, meaning that potential outcomes are exhaustive with no chance for error.
But remember, AI is based on statistics, which means probabilities, which means inherent uncertainties when using any LLM. This goes back to the fundamental understanding of the technology and how to apply it practically and strategically.
NLX strives to be a leading voice in the practical and profitable use of conversational AI applications at the enterprise level. If you want to learn more about our approach, we’re happy to chat: https://nlx.ai/contact#GenerativeAI#AI#Earnings
Waystar Royco we’re not, but we are a family business! CEO and chief product officer Andrei Papancea (right) and his brother Vlad Papancea (left) founded NLX in 2018 with close friend Peter Szerzo.
In just a few years, these bros scaled NLX into a global organization serving some of the largest and most admired companies in the world. All the growth, all the potential, none of the drama. 🚀🗺️🥰
#BornToScale#FamilyBusiness#GenerativeAI
An article in MIT Technology Review by David Rotman does an excellent job of explaining how economic growth as a result of increased productivity from Generative AI is still to come. He also talks about the power of AI to unlock major scientific advancements that could transform the economy. Article is here: https://lnkd.in/ejNgiEg2
The only bone to pick with the piece is that some of the capabilities describing a “future vision” for AI capabilities are actually quite possible today. So we would argue that the fact that these applications haven’t come to fruition is about lack of awareness rather than technical feasibility.
The example of multimodal AI being applied by a factory worker who takes a photo of a machinery problem and asks an AI assistant for solutions from the company’s operating manual, and including any regulatory guidelines, leveraging real-time data from the machine itself can be implemented in a matter of weeks using NLX today.
What’s more, the non-technical factory staff can come up with new ideas for NLX AI applications for any aspect of their work and easily implement them with minimal engineering support.
At NLX, we already knew about the difficulty in getting conversational and generative AI applications into production, which is why we designed the platform so that implementation is plug-and-play, and that the people using the applications can manage, change, and modify them with minimal to no engineering support. The people on the ground, or on the factory floor, are empowered to innovate and execute their own ideas. This approach accelerates the pace of innovation and productivity within an enterprise.
This is multimodal like you’ve never seen before. NLX is interacting with an end user setting up internet service via voice while responding to her text inputs on the screen in real time. It almost makes us wish we had a customer service request!
#MultiModalAI#GenerativeAI#CustomerExperience
As a United Airlines customer, you’re at the heart of our AI journey. ✈️
I’m headed to SXSW in 2025 to explore this new era of more personalized travel with Brett Craig and Lori Beer. But to make it on stage, we need you to cast your vote for our panel “Beyond AI Hype: Customer-Centric Tech Strategies” here: https://lnkd.in/dvtf8VtK
Your feedback and insights play a crucial role in bringing advancements like ConnectionSaver and Every Flight Has a Story to life, so I’m excited for this chance to share our vision of the future of AI.
We want to call out a specific aspect of The White House “Time is Money” initiative, which addresses the issue of AI Assistants purporting to be human to end users.
The timing is perfect, as with the continued advancement of AI voice technology, there is increased risk of consumers being led to believe they’re interacting with a real person, and unfortunately many AI tech providers are embracing this approach.
We encourage the Biden-Harris administration to ensure they’re considering voice technology as well as text-based chatbots in its efforts.
#TimeIsMoney#AIAssistants#GenerativeAI