Top 6 AI Predictions and Trends for Enterprises in 2024

To say that 2023 has been a transformative year for AI would be an understatement. We witnessed relentless innovation, surprising breakthroughs, and new AI products coming in so fast it’s been hard to keep pace.

Now, as we stand on the brink of 2024, it’s time to ponder what the next year holds for AI. If the breakneck pace of 2023 has taught us anything, it’s that nobody can predict precisely what lies ahead. But we can start with examining the top 3 trends and key drivers from this year which will continue in 2024:

1. More businesses experimenting with GenAI, but there’s a big gap to production

A Gartner survey in 1Q’23 showed that 70% of organizations were purely investigating GenAI and only 4% had gone live with GenAI solutions. But when the same questions were asked in 3Q’23, 45% were piloting or experimenting, and another 10% had gone live with GenAI solutions. This is a testament to the rapid investment being made in AI, and the pace of adoption and scale.

2. Moving beyond the hype: It's time to get real

We’re starting to see the big burst of excitement surrounding GenAI start to stabilize, and businesses and investors have started becoming practical again. Organizations will focus for AI projects with tangible business outcomes and establish methods to measure success. Similarly, VCs are now looking for strong business fundamentals and looking for things like unit economics, profitability, etc.

3.??? We're still in an AI Race

Despite the constrained economic environment, companies are racing to advance AI and tap into the opportunity areas. Incumbents are integrating AI into their existing products, while startups are innovating with reimagined experiences and products. This race is marked by rapid innovation and intense competition, with no clear winner that’s emerged yet. Each company is developing their unique strategy for sustainable growth, and 2024 will show no signs of slowing down.

?

Now, let's explore the Top 6 AI Predictions for Businesses in 2024:

1.??? Bigger Isn't Always Better: Businesses will Deploy More Smaller, Specialized Models in Production Rather than Large, General-Purpose Ones

Large, general-purpose models are great for experimentation and early prototyping, but require huge amounts of compute and storage to develop, tune, and use, and thus may be cost-prohibitive to most organizations.

As companies move to operationalize AI in 2024, they will make decisions on the models to use not based on how many parameters the models have but based on their effectiveness on domain-specific tasks and their efficiency.

Increasingly, smaller, more focused models are proving to be not only more economically viable but also superior in performance for specific tasks.

In 2024, companies will build out their own portfolio of specialized models, each fine-tuned for a specific function and minimally sized to reduce costs and boost performance. This shift will lead to a greater number of smaller, focused models being used in production deployments. ?

2.??? Multi-modal, Multi-Sensory models: The increased prevalence in everyday life and industry applications

This year, we saw advances in specialized multimodal AI systems that process and understand information across various modalities, including text, images, 3D, audio, video, and even brain activity . As these models become more sophisticated, my prediction for 2024 is that we’ll see these technologies become more integrated into everyday life.

These AI applications will be deeply context-aware and capable of complex, nuanced interactions. An example is the recently announced wearable AI pin , indicative of emerging AI-powered devices and agents that can understand and respond to a broad range of sensory inputs and, eventually, autonomously initiate actions with minimal human instruction.

2024 will be the year we see businesses use these AI systems in real-world settings, ranging from conversational interfaces to autonomous systems. Such systems will revolutionize industries such as healthcare, education, robotics, and e-commerce and transform how we interact with the world around us.

3.??? Things Like “Data” and “Security” Will Become Mainstream Again

As companies focus on operationalizing AI in 2024, they'll encounter gaps in key areas like data quality, security, privacy, and trustworthiness, which will have to be addressed to fully harness AI’s capabilities.

For instance, quality data remains paramount for any AI project. High-quality data can help build more effective models even with smaller datasets, and this data can also be better leveraged in AI applications incorporating retrieval. In 2024, we will see organizations prioritize building strong data foundations and governance.

GenAI also increases the risk of more sophisticated cyber threats. And the use of GenAI without precautions in place also open up companies to potential vulnerabilities such as model tampering, breaches, and more. Businesses will invest in proactive security measures and tools to mitigate these risks.

GenAI can also be a part of the solution. Severals tools in the market are exploring the use of GenAI to automate data analysis or enhance incident response.

4.??? Open-Source AI Models Will Become Increasingly Competitive, And Outperform Closed Models In Market Adoption

Open-Source is a big part of AI’s future. The most notable open-source AI models that emerged as strong competitors this year were Meta’s LLaMA2 70B, Mixtral 8x7B, Yi 34B, and Falcon 180B.

Open-Source model development is rapidly accelerating and these models already outperform leading closed models on select tasks. Tools like IBM’s watsonx are also catalyzing this shift by making it easier for developers to access, test, and tune these powerful models.

The recent OpenAI leadership saga was also a reminder for businesses on the risks of relying on a single AI provider and the need for business continuity. Companies are now prioritizing flexibility in their AI strategies, seeking alternatives that allow for more autonomy and control, and open-source models offer an effective path to achieve that. In fact, Forrester predicts 85% companies will expand AI with Open-Source models.

5.??? Diverging AI Regulations Will Result In More Phased AI Roll-Outs And A Temporary Slowdown In Product Development Velocity

As the AI regulatory landscape evolves, we're seeing different approaches by different governments. The EU AI Act, for example, is proposing restrictions on the purposes for which AI can be employed, placing more scrutiny on high-risk applications.

In the US, President Biden’s AI executive order took a slightly different tune, focusing on the vetting and reviewing of models and imposing restrictions and standards based on that. Canada’s AI and Data Act and India’s Digital Act seem to be leaning towards a use case driven risk categorization approach in line with the EU’s AI Act.

In 2024, companies will prepare for adhering to new standards on consumer data privacy, bias, discrimination, and cybersecurity. While the regulations are being defined, organizations such as the AI Alliance and Frontier Model Forum can help companies strategize how to self-regulate their AI solutions. But these activities will decelerate the pace of product development in the short-term. This shift will also result in companies taking more phased approaches to AI rollouts, which will increase time to value for customers.

?

6. ???A Major Insurer Will Offer An AI Hallucination Insurance Policy

As the use of GenAI proliferates, so do potential risks - with one of the most prevalent being the risk of ‘hallucinations’ or false outputs. In 2024, insurers will adapt their risk management offerings to include specialized coverage for financial losses stemming from AI model failures. This is akin to the rise of cybersecurity insurance after notable security breaches.

Many leading industry players already provide IP protections for their GenAI models or tools. However, the novel concept of hallucination insurance can offer coverage for financial damages from errors and failures in the overall AI system.

Though still an emerging concept, such insurance can also address risk management concerns from stakeholders across the board. Gartner predicts over fifty percent of large enterprises will buy some of these insurance policies if they prove practical and relevant. And they expect this to be a potentially profitable venture for insurance companies in 2024.

Closing Thoughts

As we navigate this uncharted territory, one thing we know for certain – 2024 will not be boring. So buckle up – it's going to be another fascinating trip!

As we bid farewell to a year of unprecedented change, let's also prepare to welcome a new year with open minds and a readiness to adapt to whatever comes our way.

Here's to 2024 and the continued journey of AI shaping our world in ways we've yet to fully understand!

?

Disclaimer: The opinions in this article reflect my personal views. And while they are shaped by extensive dialogues with founders, investors, business leaders, and you (my enriching LinkedIn Community), they do not reflect the views of my past, current or future employers.