Generative AI Will Impact Us All
By Neil Cornish, Business Development Manager at TD SYNNEX
When TD SYNNEX (then Tech Data) set up an AI Centre of Excellence in 2017, we knew it was the start of significant change for many industry solutions.
At the time, we worked with the likes of DeepZen, Oxford Cancer Biomarkers, and the Velindre Cancer Trust, who were all testing early ideas. We quickly saw the potential impact of trained foundation models for detecting cell types, mimicking the human voice when reading text and understanding a concerned patient’s request within a chatbot conversation and providing a consultant’s approved response. All of these had a huge amount of subject matter expertise spent on researching the problem they were trying to solve and consumed both compute and human resources to continue their development.
They all needed significant trialing before they could be used in real-world situations. The time to value was measured in years because of the need to scale, show measurable results and gain funding for follow-on developments.
The Evolution of Generative AI Technology
Fast forward six years, and so much has changed. The technology to enable all of these solutions is readily available. There are multiple machine learning models that can be trained to recognize a change or abnormality within an image with incredible accuracy. This is because of the increasing size of the datasets and their diversity, along with increased compute performance and a greater understanding of how machine learning actually works. And the time taken to train a voice model and the amount of training data required from an individual to sound authentic has reduced significantly.
The availability of computing resources in the cloud, coupled with vast amounts of public data, is something that ChatGPT has brought to light. In effect, giving their large language model years of internet data from which to learn. ChatGPT reached over one million users within a week of its launch, the fastest of any application to date, and with it, an indication of how the interest in AI has exploded over the past 12 months. We are now seeing many examples of AI-generated content all over social media, some intended to inform and, unfortunately, others to deceive.
This form of artificial intelligence, known as Generative AI, combines different machine learning techniques working on huge datasets. I decided to ask ChatGPT for its definition of Generative AI.
“Generative AI refers to a type of artificial intelligence system that is capable of creating new content, such as images, videos, music, or text, that is not explicitly programmed by a human. Instead, these systems are trained on large datasets and use machine learning techniques to generate new content that is similar to what it has seen in the training data.
Generative AI algorithms typically use deep learning models, such as Generative Adversarial Networks (GANs) or Variational Autoencoders (VAEs), to generate new content. GANs work by training two neural networks, a generator and a discriminator, in a game-like setting, where the generator tries to produce realistic content that can fool the discriminator into believing it is real. VAEs, on the other hand, use an encoder-decoder architecture to learn a compressed representation of the data and then generate new content by sampling from this learned distribution.
Generative AI has a wide range of applications, from creating realistic virtual environments for gaming and simulations to generating realistic images for medical diagnosis, to generating creative content in fields such as music and art.”
The text in the rest of this article that is not highlighted was written by me — honest!
The Impact of Generative AI on Human Careers
When asked by IDC and published in their Tech Buyers Guide to Generative AI, global execs expected to see the most significant business impact of Generative AI on software development and product design. Customer Engagement, Marketing, and Sales were listed next, and then HR.
The publication of this report has coincided with IBM announcing that they would not be recruiting for roles that in the future could be undertaken by AI.
What roles might they be? Having spent most of my career in direct sales, I wondered what would happen if purchasing departments were replaced by Generative AI. Maybe it could improve overall efficiencies, reduce the time negotiating, ensure requirements were clear and easy to bid against, and ensure emotions and human factors did not get in the way of the correct decision.
But this might reduce the need for traditional salespeople, who have built careers on their network and have a reputation for delivering. Surely, Generative AI will want to know if you can do it, how you will do it, when you will do it and what it will cost. Where your kids go to school or which golf club you belong to won’t have any bearing on the decision.
So, I asked ChatGPT if it thought it would be replacing a purchasing department. Here is a very savvy answer. Notice the use of “near future” twice.
“Generative AI has the potential to automate some of the tasks performed by a purchasing department, such as processing orders, managing inventory, and sourcing suppliers. However, it is unlikely that it could replace an entire purchasing department in the near future, as there are many tasks that require human decision-making and interaction.
For example, purchasing departments are responsible for negotiating with suppliers, developing relationships with vendors, and making strategic purchasing decisions that align with the organization’s goals and values. These tasks require human judgment, creativity, and communication skills that are not yet replicable by generative AI systems.
Additionally, purchasing departments often work closely with other departments within an organization, such as finance, operations, and marketing, to ensure that purchases are aligned with the overall business strategy. These cross-functional relationships and collaborations also require human interaction and decision-making.
Therefore, while generative AI may automate some of the routine and administrative tasks performed by a purchasing department, it is unlikely that it could replace the entire department in the near future. Instead, generative AI could be used to augment the work of purchasing departments, allowing human workers to focus on more strategic and value-added tasks.”
It would appear the purchasing department is safe for the “near future.” It is also interesting that it understood the wider role of purchasing and explained why human participation was required. But what about the sales team? Could Generative AI replace the sales department?
Below is the response. Once again, the term “near future” features, and we have another “not yet,” which suggests something, at least according to Generative AI, is not far away.
“Generative AI has the potential to automate some of the tasks performed by a sales team, such as lead generation, lead scoring, and customer segmentation. However, it is unlikely that it could replace an entire sales team in the near future, as there are many tasks that require human interaction and decision-making.
For example, sales teams are responsible for building relationships with customers, understanding their needs and preferences, and providing personalized solutions and recommendations. These tasks require human empathy, creativity, and communication skills that are not yet replicable by generative AI systems.
Additionally, sales teams often work closely with other departments within an organization, such as marketing, product development, and customer service, to ensure that the products or services offered align with the customer’s needs and expectations. These cross-functional relationships and collaborations also require human interaction and decision-making.
Therefore, while generative AI may automate some of the routine and administrative tasks performed by a sales team, it is unlikely that it could replace the entire team in the near future. Instead, generative AI could be used to augment the work of sales teams, allowing human workers to focus on more strategic and value-added tasks.”
So, it would appear our buying and selling roles are okay for a while, as Generative AI tries to work out how it does human interaction and empathy. Once again, it recognizes the roles that a salesperson plays.
Hypothetically Speaking, What If…
But what happens if the two AI personas work out that it’s easier if they just buy and sell to each other? Then neither human role is required.
The buyers train AI how to buy, sellers train AI how to sell, and they get on with it. Short term, there is a requirement for subject matter experts, just as there are with radiographers and cancer consultants when training their models.
But once the training is started, then the GANs will take over, and the buyer/supplier relationship will be handled with little human interaction. However, this scenario, similar to the introduction of autonomous vehicles, doesn’t have an overnight cut over.
If all cars were autonomous tomorrow, thousands of lives would be saved. It is estimated that if 10% of US vehicles were autonomous, 1000 lives would be saved each year. AI is a far better driver than most humans, and if every car knows where every other car is, then accidents are far less likely to happen. And AI doesn’t get drunk, take drugs, or fall asleep at the wheel. Our roads will continue to have a hybrid model for many years to come and lose many more lives because of it.
This hybrid model will continue to be the case for many business functions too. Human interaction with AI will increase, and for now, it will be our experience of the capabilities shown that will help us select or otherwise, a supplier. Most chatbots from banks, phone companies, and insurance companies drive us to wanting to speak with a human. When a company’s chatbot provides a quicker, better-informed, more relevant service than a human, that business will undoubtedly see revenues increase.
The speed of adoption of AI will continue to differ by industry, company size, and country. But the impact will be felt by all of us over the “near future.”
Society is not ready for the pace of change we are about to encounter. How will we detect when someone who sounds like your son and looks like your son and wants your bank details for something, like your son, isn’t your son? If you are computer literate and follow social media trends, this may not be your problem. But if you are less tech-savvy, it can be.
Education, awareness training, planning for change is now becoming of real importance for both businesses and governments. We need up-to-date policies to protect our data, our IP, and ourselves.
AI is fast becoming both the human race’s biggest opportunity and potentially our biggest risk.
