Since the arrival of large language model-powered chatbots, like OpenAI’s ChatGPT, the corporate landscape has been frantically striving to invest in and adopt generative AI.
Executives floated (I mean salivated over) the possibility that generative AI could replace a staggering number of roles throughout virtually every sector from law to content creation and entertainment. Well, just look how well that turned out. The legal backlash has, in many cases, been severe and, just over six months into the generative AI hype cycle, cracks are beginning to show.
Whether we’re talking about the ethical issues of training LLMs and image generators on the work of artists and writers without their knowledge or consent, the fact generative AI will just make stuff up sometimes, or the revelation that running something like ChatGPT consumes the energy equivalent of 33,000 US households per day, the issues with generative AI just keep mounting. Despite these issues, generative AI is monopolising the tech investment landscape, with 40% of all Silicon Valley investment in the first half of 2023 being poured into GenAI startups.
But what about the applications? Surely all these issues and all this money is going into generative AI technology for a reason, right? Surely we all learned our lesson from the Metaverse, the crypto bubble, NFTs, and streaming and… I guess we didn’t, did we?
Well, actually, there are a few, but they won’t look like the Wild West of content generation we’ve seen so far.
In the retail sector, for example, 98% of companies plan on investing in generative AI in the next 18 months, according to a new survey conducted by NVIDIA (a company with an admittedly vested interest in selling shiny new GPUs). Early examples of adoption in the sector have included personalised shopping advisors and adaptive advertising, with retailers initially testing off-the-shelf models like GPT-4 from OpenAI.
However, many retailers are recognising that the strength (and weakness) of generative AI is that you only get out what you put in. That’s why the technology is, ultimately, useless as a way to replace creative roles like writers and artists. However, as a brand communicator meticulously trained on a specific set of data with carefully updated parameters, it could be invaluable. NVIDIA’s report notes that “many are now realising the value in developing custom models trained on their proprietary data to achieve brand-appropriate tone and personalised results in a scalable, cost-effective way.”
Generative AI trained on a company’s internal and customer-facing databases, web presence, and curated information resources could conversationally recommend, educate, and explain critical information to employees, customers, and business partners effectively and consistently. In an industry where communication relies on clarity and an understanding of large quantities of information, like procurement, the applications suddenly start to look a lot more appealing.
Chatbots and negotiation bots trained to converse with suppliers, programmed with company approved negotiation tactics and the latest pricing information, could automate a great deal of complexity out of the Source to Pay process.
I think the looming issue is the impact of generative AI adoption on a company’s Scope 3 emissions, as 2024 will unquestionably be defined by greater scrutiny on these sources of pollution. However, it seems that however many issues the more widely known aspects of generative AI have, the technology itself could still have a role within the procurement function of the near future.
Does it justify all the investment, hype, and endless industry media thinkpieces? I guess only time will tell.
By Harry Menear
