Welcome to the first working Monday of 2024. AI may feel sooo 2023, given the explosion of interest generated by the public launch of ChatGPT. But interest is unlikely to dwindle in 2024, even if our focus moves on from “write a resignation letter to my boss in the style of a Hamilton hip hop song” to “help me define my target market segments, create personas and write my marketing plan…”

So, which Generative AI trends are worth keeping an eye on in 2024? Here are just a few thoughts. Tell me what I’ve left out:

Capability

• Much of 2023 saw OpenAI ’s competitors playing catch up with ChatGPT 3.5, and then GPT-4. None of them quite made it, but Google claims its soon-to-be-launched Ultra version of its ChatGPT alternative, #Gemini, will beat #GPT-4 in some areas. That’s if OpenAI’s doesn’t launch GPT-4.5 first…
• Why does it matter? At the launch of GPT-4 OpenAI said that it’s “more reliable, creative, and able to handle much more nuanced instructions than GPT-3.5”. In practice, it became “multimodal” (in other words, it could handle visual, written and spoken instructions and responses), users could expand and tailor its use with the introduction of plugins and then GPTs (tailored versions of ChatGPT that users can create without any coding knowledge) and it came with DALL·E 3 – OpenAI’s image generator. These are just some of the ways the improvements to GPT increased ChatGPT’s capabilities in 2023. Expect more in 2024.

• The drive for more capabilities more quickly will also fuel the debate as to which is better: open source language model development vs closed source. In other words, community collaboration vs proprietary focus. The closed source crowd will continue to be led by OpenAI – ironic given their name. In the open source world, Meta’s soon to be launched #LLaMA-3 appears to have real competition in engaging developers in the form of French company Mistral Solutions Pvt. Ltd ’s catchily named “Mixtral 8x7B”. Curiously, Apple, not a company famous for its open sourced approach, released #Ferret (yes, you read that right) a generative AI model for vision-language tasks, as open source last October. Expect to see it incorporated into a turbo-charged version of Siri that we all wanted to see launched several years ago.

Adoption

• While companies will continue to be concerned about inaccuracy, confidentiality, security and copyright issues, adoption of Generative AI tools at work will accelerate, propelled by personal use, much as the iPhone’s initial enterprise adoption was driven by employees rather than company policy.

• Adoption of Generative AI across functions and sectors will broaden. Marketing and Sales teams largely led the charge in 2023. “Industries relying most heavily on knowledge work are likely to see more disruption—and potentially reap more value” says McKinsey & Company in “The state of AI in 2023: Generative AI’s Breakout” report. Expect to see an acceleration in adoption especially in banking, pharmaceuticals, medical, legal and accountancy sectors. But, also in sectors that have been slower to the table, such as retail and manufacturing, as the value in more efficiently and effectively making use of customer insights to provide “hyper-customised customer experiences” becomes more accessible.

Integration

• Key to adoption will be an increased focus on integration. On the device front, we’re seeing innovations like Google ’s Nano version of multimodal model, #Gemini – designed for “on device experiences” aka mobile devices, the surge in marketing of laptops as “AI PCs” complete with “made for AI” chips by the likes of NVIDIA . We will also see more new devices, along the lines of Humane’s AI pin, and the Rewind Pendant, as different parts of our body are deemed to be fit for purpose as ai containers. We can also expect big things (or maybe small and very intuitive to use things) in 2024 from the collaboration between Apple’s ex design guru, Jony Ive’s #LoveFrom and Sam Altman’s OpenAI. At the end of the year Reuters reported that Tang Tan, “who led the design for the iPhone and Apple Watch” is set to come on board in February.

• We will also see more integration of generative AI into the enterprise systems that we already use. Leading the way is Microsoft ’s #Copilot integration into Microsoft 365 Enterprise and Windows 11, and its recently announced new, dedicated Copilot key to be added to the Windows keyboard – the keyboard’s first change in 30 years.

• Integration into the workflow will be propelled by the adoption of AI tools or agents. OpenAI’s launch of GPTs – tailored versions of ChatGPT for a specific use or task – is an example of those tools, and an “App Store” like platform to share and sell them is due from OpenAI this month. Google is rumoured to be launching a similar tool, under the Gemini brand, later this year.

• We will also see more use made of Generative AI in business processes and decision making. More businesses will start using AI for strategic decision-making processes, such as analysing market trends, forecasting, and providing insights for high-level strategy and operational decisions, as well as recruitment and training, regulatory compliance and helping to adopt and embed better ethical and sustainable practices.

From Images to Video

• If the latter half of 2023 saw major advances in image generation – led by OpenAI’s DALL·E, Midjourney , Craiyon🖍 and #DreamStudio by Stability AI – then expect to see the same for video in 2024. This will likely be led by #Pika Labs, #Midjourney, Runway, Stable Video Diffusion, Synthesia and others. Image generation will see further adoption as DALL·E 3 has been integrated into Chat GPT Plus, Midjourney becomes accessible via a website rather than just Discord and more people make greater use image generators that are integrated into the design platform they already use, such as the Canva AI Image Generator and Adobe #Firefly.

Loneliness and Companions

• As Prof Scott Galloway has so forcefully argued, one of the biggest malaises to hit Western societies is loneliness. The Survey Center on American Life reports that in a US-based survey comparing 1990 and 2021 the percentage of people claiming they have less than three close friends doubled from 16% to 32%. The same survey reported that the percentage of people who claimed to have no close friends at all rose from 3% to 12%.

• Perhaps not totally disconnected, the Character.AI website was usually to be found in the top ten AI websites by use, in the 2023. Character AI lets you create “characters” and talk to them. In other words, #AIchatbot companions. Expect to see AI friends become more prevalent, as a synthetic antidote to loneliness, providing digital companionship.

Legal, regulated and honest

• Copyright, privacy, and regulation are key concerns and will remain so in 2024, as countries grapple with finding a balance between providing a stable, ethically acceptable environments for the development of AI technology without impeding the competitiveness of their companies. All while trying to keep up with the fastest-developing technology ever. The EU’s proposed AI Act is a case in point. It emphasises a risk-based approach, and has been several years in the making, which means it was conceived before the explosion in Generative AI, and President Macron, with an eye on the French company Mistral, is concerned that it will prove to be too restrictive in the face of foreign companies.

• On the copyright front, the case The New York Times has brought against OpenAI will, coupled with the report that Apple is spending approximately $50 million to license content with key international media bodies, including IAC , Condé Nast , and NBC News , will help to determine the value of original content used to train AI models.

• A US Presidential election, as well as national elections in a number of countries will bring the dangers of AI created “deep fakes”, sharply into focus in 2024. Developments like the ability of AI tools to pass the Capchta test, highlighted by Professor Ethan Mollick at The Wharton School, highlight the very broad range of such challenges.

• I’m not convinced 2024 will provide the answers, but the commercial opportunities for the “verification” market will be substantial. The recent debates on defining and detecting plagiarism in US universities is an obvious example. And, it almost goes without saying, but the value of trusted brands will only increase.

Medicine and Health

• I hope we see more AI-fuelled health and medical breakthroughs in 2024, for example to identify the earliest signs of various cancers, vaccine breakthroughs, predict cardiac arrest, mitigate the risk of diseases like HIV, uncover genetic predictors of diseases such as Alzheimer’s and provide basic, affordable health care to areas of the world that still lack it. Bill Gates’s blog post, “The road ahead reaches a turning point in 2024″, provided a reasonably optimistic vision. I hope he’s right.