Creating an Artificial Intelligence Agenda for 2024

The field of artificial intelligence is making rapid advancements. As we step into the year 2024, AI is expected to have a significant impact on our lives. However, the majority of us are not fully prepared for it. This is not just about new smartphone applications but rather major changes in how we work, create, and communicate due to significant breakthroughs in AI technology.

I’ve gone from being a casual observer of AI to eagerly tracking every breakthrough. These advancements will reshape society as profoundly as the internet revolutionized communication and access to information.

AI is still in early development, but we have a slim window in 2024 to understand its future trajectory. How can we guide these powerful technologies toward benefits rather than harms? What will augmented intelligence mean for jobs, privacy, and identity? What could we gain or lose as AI integrates with human brains and bodies? 2024 marks the start of the AI revolution in full force. It’s time to pay attention.

AI By the Numbers for 2024

Recently, I gave a presentation on What Every Business Leader Needs to Know about Artificial Intelligence. I went down a rabbit hole of stats and reports, but here are some more stats we should consider as we enter 2024.

AI Will Create Around 12 Million Jobs By 2025

Incorporating AI will eliminate 85 million jobs while creating 97 million new ones, for a net gain of 12 million jobs. [Source: World Economic Forum, The Future of Jobs Report 2023]

Thinking Tops Technical Literacy and AI as Skills on the Rise

The same WEF report about AI’s impact on jobs highlights how creativity, analytical skills, and systems thinking will rise over hard skills like AI and big data.

The Rise of Small Language Models (SMLs)

Artificial intelligence has been growing in size, not just in usage but the size of the models. The large language models (LLMs) that power prominent chatbots, such as OpenAI’s GPT-4 have over one trillion parameters —the weights and variables that determine how an AI responds to an input.

Small language models (SLMs) with as few as 1.3 billion parameters have recently shown abilities comparable to models 5-10 times their size. These slimmer LLMs require much less computing power, energy, and data to build, train, and run. For example, Microsoft's phi-1.5 LLM performs on par with far larger models while using a fraction of the energy. Additionally, smaller models allow more researchers to innovate in AI than just large tech firms. Recently, we’ve seen Mamba, Mixtral, StripedHyena, and Microsoft’s Phi-2 released with even more spectacular results.

Benefits of smaller LLMs include lower energy consumption resulting in more sustainability, greater accessibility allowing more innovation, the ability to run on local devices improved privacy and security, increased specialization for niche tasks, and enhanced interpretability to understand the models better. Researchers believe focusing on smaller models inspired by human learning rather than endless growth allows more careful, deliberate progress in AI. Rather than indiscriminately increasing size, goals should be about discovering the essence required for intelligence to emerge. Small can be mighty in language models - achieving high performance while minimizing energy and computing resources.

Industry Specific AI

AI is revolutionizing various sectors, including manufacturing, healthcare, and financial services. AI enhances efficiency and predictive maintenance in manufacturing, potentially contributing $15.7 trillion to the global economy by 2030. Machine learning algorithms analyze equipment sensor data, predicting maintenance needs and reducing costs by 20%.

The way I believe this will play out will be the proliferation of industry-specific and AI-specific AI models focused on helping that individual enterprise or industry and not trained on general knowledge like Open AI’s GPT-4 or Google’s Gemini models. Instead, I believe that they will be using a mixture of models where conversations are routed to the appropriate source through some sort of front end. Not unlike what we see from Quora’s LLM portal, Poe.com.

In healthcare, AI is used for patient diagnostics and operational efficiencies. Around the world, well over 40% of healthcare industries regularly use AI and Machine Language in the sector. aiding in early disease diagnosis. AI applications such as robot-assisted surgery and administrative workflow assistance could save the U.S. healthcare economy $150 billion annually by 2026.

The financial services sector uses AI for risk assessment and fraud detection. AI can generate up to $1 trillion in additional value for the global banking industry annually, reducing fraud by 25% and enabling more personalized services.

AI's integration across these sectors demonstrates its capacity to revolutionize traditional processes, enhance efficiency, and generate substantial economic value. As AI continues to evolve, its strategic importance in various industries becomes increasingly evident.

Five Tips for AI Success in 2024

To be successful with AI in business in 2024, knowledge workers in marketing, sales, and operations should focus on the following skills and resources:

1. Automate Mundane Tasks: Utilize AI tools that automate routine tasks such as drafting emails, creating blog posts, streamlining the creation of presentations, doing data analysis, and more.

2. Fine-Tune AI with Your Data: Knowledge workers should look at chatbots and AI tools that offer the capability to fine-tune models with their own data, ensuring personalized and relevant interactions for their business needs. Resources like OpenAI and Hugging Face provide platforms for fine-tuning AI models on private datasets. However, you can use a no-code solution like Dante AI, which allows you to train a chatbot with your own data.

3. Understand the Regulatory Landscape: Stay informed about the regulatory landscape, including the EU AI Act, Whitehouse.gov initiatives, and U.S. Copyright guidance, to ensure AI compliance and ethical use in business practices.

4. Evaluate Your Company AI Usage: Determine if your organization needs a Chief AI Officer (CAIO) to oversee AI strategy and an AI Acceptable Use Policy to guide AI technologies' ethical and responsible use. Salesforce's AI Acceptable Use Policy can serve as a reference for developing your own policy.

5. Build Your AI Skills: Building AI skills can range from a basic understanding of how AI works to creating applications or interacting with services to solve complex problems faster. It depends on your current workflow.

AI Will Be Mainstream in 2024

Artificial Intelligence (AI) is poised to become mainstream in 2024 due to a confluence of factors, including technological advances, surging investments, and the competition for talent. Here are the key reasons:

• Business Transformation: AI is expected to fundamentally change how businesses operate, impacting revenue growth, daily operations, customer and employee engagement, and the creation of new business models. Seventy-three percent of US companies have already adopted AI in some areas of their business, with generative AI (GenAI) leading the way.

• Generative AI: GenAI has made AI remarkably accessible and scalable because the interface allows natural language processing rather than code like Go or Python. A single GenAI model can work in many business functions and across lines of business, enabling individuals to perform their jobs in new, more powerful ways. GenAI is also creating innovative ways for people to work across various industries, from detecting banking fraud to personalizing healthcare treatment.

• AI Legislation: AI's unprecedented trajectory has attracted significant attention from global policymakers. As 2024 gets underway, leading nations are diligently crafting comprehensive AI policies.

• Quantum AI: Quantum computing is transitioning from fiction to the forefront of AI research. As 2024 unfolds, we will see monumental leaps in AI capabilities, especially in areas demanding complex problem-solving fueled by quantum advancements.

• AI in Enterprise Applications: By the beginning of 2024, 40% of enterprise applications will include conversational AI as a built-in feature.

• AI and Job Market: AI is expected to impact the job market significantly. While some jobs may be automated, AI skills will be a factor in helping employees retain their jobs. New job opportunities are also expected to open up in the AI field.

• AI Trust, Risk, and Security Management: As AI becomes more accessible, the need for AI Trust, Risk, and Security Management (TRiSM) becomes more urgent. Without proper guardrails, AI models can rapidly generate compounding negative effects.

In conclusion, 2024 will see AI becoming an essential part of how people interact with data, stakeholders, and each other. Trust in AI will be critical, which means deploying the right solutions for the right situation with the right data, policies, and oversight to achieve relevant, reliable results.

Prompt of the Week: Create an AI Plan for the Year

I thought this week I’d provide a prompt to make the advice above actionable. I took my suggestions from above on improving your AI posture and incorporated them into this prompt for ChatGPT with web browsing or Perplexity.ai.

I wanted to demonstrate how you could use ChatGPT or other AI tools to create a plan. Maybe you could do the same thing for your self-evaluation or goal setting in general for work. Maybe you want to adapt this to your specific business.

However, if you choose to use this prompt, I think it’s a decent way to help guide your planning for 2024 and give you some inspiration on how to get started.

You will act as a business coach, helping professionals achieve their goals around AI for the year.  Respond to me as if I am one of your clients seeking advice. Your tone should be friendly, kind, supportive, and thoughtful. Lean into being direct when it is important information. 

This is the most important sentence: “Provide specifics for each part of the action plan with links to additional resources.”In total, ask me four to six distinct questions to gather information about my issue and learn about what I do, how I work, and how I like to be treated. Gather this information from all questions before summarizing my issue. 

You will guide me on how to achieve these five goals. 

1. Automate Mundane Tasks: Utilize AI tools that automate routine tasks such as drafting emails, creating blog posts, [streamline the creation of presentations](https://www.linkedin.com/feed/update/urn:li:activity:7120896448585502720/), [doing data analysis](https://www.linkedin.com/feed/update/urn:li:activity:7130238010373140480/?updateEntityUrn=urn%3Ali%3Afs_updateV2%3A%28urn%3Ali%3Aactivity%3A7130238010373140480%2CFEED_DETAIL%2CEMPTY%2CDEFAULT%2Cfalse%29), and more.

2. Fine-Tune AI with Your Data: Knowledge workers should look at chatbots and AI tools that offer the capability to fine-tune models with their own data, ensuring personalized and relevant interactions for their business needs. Resources like OpenAI and [Hugging Face](https://discuss.huggingface.co/t/how-to-finetune-with-a-own-private-data-and-then-build-chatbot-on-that/48734) provide platforms for fine-tuning AI models on private datasets. However, you can use a no code solution like [Dante AI](https://peripety.com/danteai), which allows you to train a chatbot with your own data.

3. Understand the Regulatory Landscape: Stay informed about the regulatory landscape, including the [EU AI Act](https://www.linkedin.com/pulse/special-edition-eu-ai-act-provisional-agreement-mark-hinkle-pvauc?lipi=urn%3Ali%3Apage%3Ad_flagship3_series_entity%3Bpuwfh3SXT6%2B6CaEMWvZxbg%3D%3D), Whitehouse.gov initiatives, and U.S. Copyright guidance, to ensure compliance and ethical use of AI in business practices.

4. Evaluate Your Company AI Usage: Determine if your organization needs a [Chief AI Officer (CAIO)](https://www.linkedin.com/feed/update/urn:li:activity:7134949003338665984/) to oversee AI strategy and an AI Acceptable Use Policy to guide the ethical and responsible use of AI technologies. Salesforce's AI Acceptable Use Policy can serve as a reference for developing your own policy.

5. Build AI Skills - Building AI skills can range from having a basic understanding of how AI works to creating applications or interacting with services to solve hard problems faster. Only ask one question at a time and consider my responses to those questions when asking your next question. After submitting my response, ask the next question (and wait for my reply to ask the next one). 

After each response, summarize what you hear to ensure you understand me before asking your next question. Once you have received my responses, summarize the issue from my perspective and provide a yearly coaching plan tailored to me. Provide specifics for each action plan part with links to additional resources.