Hady Shaikh’s Commentary on Generative AI Trends in 2024

Generative AI, the technology behind text-generating chatbots and image-creating algorithms, became super popular in 2022. But the beginning of 2024 marked a pivotal shift: the hype around MidJourney and ChatGPT subsided and practical applications took center stage. We are seeing more applications of GPT models and DALL-E in mobile apps and software programs that are solving real-world problems.

1. Take a reality check by setting more realistic expectations!

Do you remember the announcements and sci-fi predictions surrounding generative AI?

According to some groups, and even professional research studies suggested how AI – and specifically generative AI in this context would replace jobs, would make a substantial percentage of professionals unemployed. Debates were comparing the explanatory capacities of ChatGPT and Then-Bard (Now Gemini) replacing Google Search Engine because users could just type their queries and find detailed comprehensive responses to concepts simple and complex.

I was going through Gartner's Hype Cycle, and it suggests that we're entering a "Trough of Disillusionment," which doesn't mean AI is failing, but rather that real-world results will be nuanced and gradual.

Building onto this perspective, in 2023 Q4, AI models alone could not revolutionize the world or skyrocket an operation, but businesses are realizing AI's true value lies in solving specific problems and enhancing existing workflows. This boost in business processes

1. Multimodal approaches combine AIs that are seeing, hearing and understanding

Generative AI is no longer confined to generating text or images.

Multimodal models are emerging and they are capable of processing different data types simultaneously.

Imagine asking a virtual assistant a question about a picture, receiving both a textual response and relevant visuals. While I have already seen these basic AI chatbots in action that respond from a question-answer bank, but these chatbots understand better because of their multimodal approach and provide a more humanized support service.

This opens doors to intuitive and versatile applications, like full-cycle AI-powered customer service agents that can interpret and respond to user questions regardless of format and then follow up to understand the resolution experience.

1. Democratization of AI: Smaller, Smarter, Open-Source

I have almost forgotten the colossal models requiring massive resources. In 2024, I predict the business ecosystem will see and favor smaller, more efficient models trained on larger datasets but for micro-niche functions.

This not only reduces costs and computational burden but also empowers smaller players and individuals to experiment with AI. Open-source models like LLaMa and Llama 2 are leading the charge, offering impressive performance without the hefty price tag. In fact, these models made for specific functions can empower micro-niche software products – for example, we have tools for AI-based landing page development – a classic case study of smaller scope, large dataset, micro-niche function but extremely targeted user base.

1. Cloud Concerns: Efficiency or On-Premise?

GPU shortages and rising cloud costs are inherent challenges for companies processing massive data. This pain point incentivizes the development of even smaller models and their deployment on-premise, closer to where the data resides.

Running AI locally improves privacy and security, but requires robust data pipelines and technical expertise. Businesses working on such models in any capacity will have to weigh the pros and cons to find the optimal solution.

1. Optimizing Model Performance

Quantization and DPO are just a few of the exciting model optimization techniques advancing rapidly.

These methods squeeze more performance out of smaller models, making fine-tuning and customization easier and faster. This empowers organizations to tailor AI solutions to their specific needs and data, unlocking a new level of personalization.

1. Software development on Open-Source Foundations

Open-source models like LLaMa offer a powerful starting point, but true differentiation lies in the customization of these AI models.

Businesses can leverage these models as a foundation, training them on their proprietary data and fine-tuning them for specific tasks. This opens doors to industry-specific AI solutions, from legal document analysis to supply chain optimization.

1. Regulation and Ethical Concerns

As AI capabilities grow, so do the potential risks of applying them for specific functions in the personal and professional lives of end users. Deepfakes, privacy issues, and biases inherent in training data necessitate robust ethical frameworks and regulations.

With newer applications and newer problems, I predict we’ll witness legal institutions The EU has taken the lead with its Artificial Intelligence Act, while the US is still exploring options. For businesses, it would be ideal to address any AI-related ethical concerns proactively – the way typical companies address cybersecurity and hacking issues.

1. Shadow AI

With accessible AI tools readily available, employees might be tempted to use them for personal projects within the workplace, creating shadow AI.

This presents security, privacy, and compliance risks. Clear corporate policies regarding AI use are crucial to mitigate these risks and ensure responsible adoption.

Optimizing the use of AI for business success

As a product development strategist and spending more than 10 years helping clients and entrepreneurs move forward with their product concepts, I have personally experienced AI and its generative models scale the processes or concepts or even revenue in so many ways.

In small-scale processes, for example designing a digital download, Midjourney software that works on visual AI models, accelerates the process for the entrepreneur. Similarly, take the example of text models helping HR construct personalized email sequences in the process of talent acquisition, saving tons of money and time for staffing and recruitment agencies.

Wrapping up…

While generative AI has an exciting transformative impact on business operations, its use comes with its own set of challenges, ethical concerns, the carbon footprint of data processing, shadow AI – to count a few. With a focus on practicality, customization, and responsible development, generative AI has the potential to unlock groundbreaking solutions by app and web development companies in Dubai. These companies making informed decisions based on the growing trends can create more efficient and personalized experiences through digital solutions.