Monday, October 30th, 2023
Ten AI Trends for 2024: Insights from EXIT83
Ten AI Trends for 2024: Insights from EXIT83
At EXIT83 Consulting, we spend all day thinking about how AI will change our world. In this post, I pulled out the special AI crystal ball to offer a forecast on the AI landscape for the upcoming year.
1. Rise of Multi-Agent Automation by Q4. With innovative tools like LangChain and AutoGen making waves, platforms such as Auto-GPT and BabyAGI are positioned to be front-runners for complex, layered AI implementations. Mastery in custom agent development is likely to become a significant differentiator for solution providers by year-end.
2. Exponential Increase in Data Collection from Everyday Objects. As platform providers understand what can be accomplished with AI, the breadth of data harvested from everyday objects will increase. Your vacuum, refrigerator, toothbrush, even toasters – literally everything will be collecting data about the world you live in. These devices may be connected to your smart home system or simply push data directly to the cloud. Of course smart devices have been around for years, but the velocity by which the number of simple devices will become “smart” will increase substantially.
Manufacturers of these devices will sell access to this data and provide insights about you via a subscription.
3. Demystifying AI for the Masses. A large segment of the population remains uninformed about the true capabilities of AI. With the terms "model" and "fine-tuning" becoming ubiquitous, a broader audience will realize that AI's potential extends beyond voice-activated assistants and ChatGPT.
4. Integration of Generative AI into Mainstream Tools. Prominent platforms like Microsoft Copilot, Adobe suite, and Apple Photos are only the beginning. The list of AI-integrated tools is set to grow.
5. History has a way of repeating itself. AI adoption will mimic the previous internet trend lines, but more so.
“I assign the Internet the highest level of importance…. Every product will a) be on the Internet, b) be able to deal with the fact it is being used on the Internet and c) be improved by the fact that it can take advantage of Internet services.” – Bill Gates, the infamous 1995 “Internet Memo”
Now replace the word “internet” with the word “AI”, and you essentially have Satya Nadella’s quote from the most recent Microsoft earnings call - “We are rapidly infusing AI across every layer of the tech stack and for every role and business process to drive productivity gains for our customers.”
According to a 2022 report by McKinsey & Company, the adoption of AI is still in the early stages, with only 20% of companies reporting that they are using AI in some form. However, the adoption of AI is growing rapidly, with 40% of companies planning to adopt AI in the next two years.
This compares to 1998, where it is estimated that 20% of commercial software had internet features baked into it, however it took an additional 8 years to hit that 40% number.
6. You get a custom model. You get a custom model. Everyone gets a custom model. 2024 might be marked by the surge in custom model training to address unique enterprise challenges. As conventional software development approaches saturation, AI and data engineering will emerge as the primary investment areas.
While general education around AI increases, the ability to see an AI based solution to enterprise problems will increase. Many of these problems require unique solutions, and thus 2024 will see a big rise in the training of custom models to address these problems.
7. Era of Detective AI. The proficiency of Generative AI will necessitate the rise of Detective AI – models trained to discern AI-generated content. These models will detect when AI has been utilized to create new text, images, and audio. Generative AI will become so good at mimicking humans, that only trained AI models will be able to notice the subtle differences. This will begin an arms race between Generative AI and Detective AI.
8. SaaS Companies Awakening to Data Potential. As 2023 comes to a close, it is estimated that about 25% of all Saas companies have realized how to unlock the treasure trove of insights from the years worth of data they have been collecting. This is due to an insufficient supply of engineers trained in data engineering and machine learning. This is also due to lack of general understanding of what AI can actually do beyond ChatGPT. As that supply of engineers and understanding increases, many more Saas companies will discover how to unlock the gold mine found in their data. Adoption won’t reach 100%, but it is likely to increase to at least 60%.
9. Governments will begin to regulate AI usage. Europe continues to lead the world in tech regulation. With respect to AI, that will be no different. With the divisiveness of the US government, it’s clear that the US will be unable to drive poignant legislation that addresses the key issues of AI. I’m talking specifically about job creation (and loss), copywrite infringement, IP ownership, identification theft (and spoofing), and the myriad of social implications that have been created. Europe won’t be able to address all of these issues, but I do believe that data ownership and licensing will be addressed by the EU over the next 12 months. New regulations will provide a way for creators to sue violators like Google and OpenAI and establish ways for these same companies to legally license data from those same people.
10. Cybersecurity malfeasance and threat mitigation will be big winners. The ever-growing arms race between Black Hat hackers and security researchers, just grew by two to three orders of magnitude. Take for example, how quickly a software engineer can use Copilot to develop a new website. Now realize that same LLM, fine-tuned with years of exploit data, can be turbocharged to develop some wild new exploit technology that normally would have taken years to develop. As security firms see the impending mushroom cloud of exploits, they will develop AI powered counter measures to combat these new technologies – thus taking the security arms race to a whole new level.
Enterprises will have no choice but to hire security experts who are fluent in AI if they want to have any chance of maintaining security. This will increase the need for security research, and the cost to hire these experts will skyrocket.
Honorable mention - There will be a bevy of new jobs that are entirely focused on AI innovations.
- CAIO – Chief AI Officer, a C-level executive that provides strategic oversight of how AI is leveraged in an organization.
- AI Compliance Officer – a risk management official that assures that the model or LLM used in your line of business complies with government regulations.
- IP Forensic Analyst – a sleuth that can dig into an AI model, either LLM or custom model, to determine if any specific IP was used in the training of the model.
- Prompt Engineer – this sort of exists today, but its numbers will increase in size. This unique engineer is someone who can orchestrate an LLM to provide answers to very complex questions. It’s requiring one part psychologist, and one part engineering degree.
- Digital Cloner – Now that LLMs like GPT4 can create sound, video, and text, it seems very reasonable that a scrapy engineer can develop a digital representation of a person. Think of being able to interact with this digital person in real time using an Apple Vision Pro or Oculus.
These are just a few of the AI trends we can expect to see in 2024. What did I get right? What is missing? What are some new AI jobs that you can think of?
How will AI change your world?