Transfix Experts Weigh In On Its Place in the Freight Industry
ChatGPT has entered the…chat! But for how long? Well, if you ask your coworkers in marketing, they’re likely to roll their eyes but others are hoping to use its advantages to help boost their daily efficiency. Its intelligence is positioned to be the perfect aid to those in customer service, research, and engineering.
OpenAI’s ChatGPT is like Google on steroids. You enter a prompt with details as specific as the format you want for your response, length of response, type of response, and even what tone you want. From there it accumulates the (alleged) best response based on researching information from all corners of the internet and talks to you as if you’re in, well, a private chat with a robot – a la “Black Mirror” style. Think less weblinks, more a large piece of text with the answer you’re looking for.
According to PCMag, it’s just smart enough to pass an MBA exam (barely). Yet, the software has limited knowledge of the world beyond 2021. And a lot can happen in a year…
So, do ChatGPT and other Large Language Models have a place in the freight industry? We polled our audience on LinkedIn and 73% are in support of it, but the Transfix team has been engaged in this healthy freight debate for nearly a decade–automating mundane manual processes in the supply chain. See what our experts think below.
- A Gamechanger for Communication Between Shippers and Carriers
On the driver side, rather than fill out long checklists of which days they are willing to accept loads, drivers will be able to say “get me home for date night or football on Friday” and let generative AI handle the rest. And while out on the road, it could be beneficial with accident detection, and driving routes that automatically adapt to hazards, traffic, and weather. What I think might be a game-changer is its impending ability to explain tradeoffs between route options in plain English (or Spanish or Punjabi–some of the most commonly spoken native languages in trucking) to drivers. For shippers, generative AI has the potential to supercharge concierge-type interactions when dispatching shipments, assigning carriers, presenting location updates, and choices for redirecting shipments when there is a delay. Functionality could evolve to presenting the most optimal tradeoffs between carriers based on greenhouse gas emissions, playing on a company’s investment in sustainability. – Ayeh Bandeh-Ahmadi, Chief Economist
- Generative AI Has Its Limitations Without Human Partnership
I would never bet against tech in the long run but I’m skeptical about the ability of ChatGPT (CGPT) to completely displace humans. Large Language Models like CGPT have garnered so much buzz because all it takes is opening a web browser to interact with the technology. This has differentiated it from other notable tech like blockchain which lacked a tangible way for the general public to appreciate its sophistication. CGPT conversations are surprisingly cogent, diverse in structure, and, depending on the topic, amazingly substantive in the way it answers questions. However, my skepticism is two-fold. Firstly, CGPT is limited in its ability to generalize beyond what has already been said or written and exposed to its training step. Its power lies in combining information from thousands of sources into a conversational response, not in deriving fresh opinions or exploring beyond the frontiers of currently disseminated knowledge. Secondly, I think the diffusion of CGPT will shift the focus of the human from collecting information to being a responsible steward of the model and its output. Technology tends to imbue us with a false sense of confidence, assuming that technological innovation is synonymous with infallibility. This new technology will be most effective in collaborative settings with human experts, empowering professionals to make the right decision with more complete information while allowing room in the process for discretion and intervention as nuance and gravity of the decision grows.
– Fil Piasevoli, Staff Data Scientist
- Your Very Own Analyst of “What Ifs” At Your Fingertips
While limited in its ability today, it could very well be an ‘analyst at your fingertips’ with the ability to query and perform ad hoc analysis on your data to evaluate how your business is doing or how service is looking with the kind of specificity that could help executives or business owners drill into a specific event. The functionality has the potential to perform projections and simulations on “what if” scenarios. What’s more, the conversational interface through both speech and text (e.g., across mobile and web) could help carriers or shippers looking for information on route updates, delivery status, appointment window confirmation or changes, and provide ad-hoc analysis of impacts of changes to service metrics, accessorials, and more. – Tony Tzeng, Chief Product Officer
- A Future Fraud Detective At Its Finest
Generative AI tools could be used to detect fraudulent activities such as fake invoicing, over-billing, double-booking, and other types of fraudulent activities in the trucking industry through the use of fraud detection techniques. This technology has already been implemented and successfully used in the banking industry, where it has proved to be effective in detecting and preventing financial fraud. – Ben Volkwyn, Senior Manager of Data Analytics