As 2024 begins, I wish to set out some thoughts on what we might expect to come with the continued dominance of AI chatter in the ether. The tech titans of Silicon Valley are all in, clearly betting their chips on the dream of generative AI transforming into a goldmine. This will happen, but when? The buzz is about tailor-made chatbots (agents or GPTs), poised to turn anyone into an AI wizard, coding skills optional. No code, low code, now it’s even easier with AI. I anticipate this likely happening this year, but will bleed into next for mass adoption. We need tools that are extraordinarily easy to engage with and create workflows. I think the largest platforms will succeed here, as will organizations with APIs that allow their data to be connected to the workflows. On the more progressive front, OpenAI is already sprinting ahead for their “app store”. I expect them to unveil their GPT app bazaar any day now.
On the #GenAI front, keep an eye out for the next wave of positive creativity and scary challenges: AI-generated videos, an avalanche of AI-fueled election misinformation, and multitasking robots. Speaking of AI-generated videos – I’ll be a judge at the upcoming Dubai AI Film Festival (#AIFF) in February and March. Keep a look out for more information on this soon! The AI video creation has improved remarkably even in the last three months since reviewing some of the films. If you have not played with Pika yet, sign-up and do so!
2024 is also shaping up to be a landmark year for AI governance globally. Following the European Union's pioneering AI legislation in 2023, the U.S. and China have also joined the regulatory dance, each with their own steps. You can guess which might be more restrictive. This year, the U.S. will likely see regulation moving from information gathering and discussion, to some crystallized actions. The primary areas of concern for many, besides the existential risk, are copyright, bias, and the high barrier to entry with cloud processing. That last point will be interesting to see develop. I will have a podcast on this and the related topic of open-source AI in the coming weeks with my guests from Booz Allen.
One other area to think about in 2024 is security with AI. Large Language Models (LLMs), like those powering ChatGPT, are alarmingly hackable. Take AI assistants with web access: they're particularly vulnerable to "indirect prompt injection," a backdoor for hackers to hijack these AI tools for phishing or scams. What I expect to happen is far greater use of these tools for scamming. As Sam Altman has stated, we are rapidly approaching a time when AI will be amazingly persuasive. If you combine that skill with the ability to swindle people electronically at breakneck speed – the epic disaster of new scams is on the horizon.
In this landscape where AI intertwines increasingly with our daily digital fabric, its flaws notwithstanding, it is vital to maintain a vigilant, discerning stance towards AI. As regulations evolve to keep pace, our collective understanding and critical examination of AI will be crucial in shaping its role in our lives and societies.