ChatGPT turns one

We blogged about AI being the 4th industrial revolution in June of this year, the pace of the technology at that point was slightly numbing, but it still felt that we were in the foothills of discovery. Fast forward to mid-November and it looks as though there has been an explosion in the deployment of generative AI tools as a regular part of day-to-day business functions. We were at the Recorded Future Predict conference in London during mid-November, where a show of hands from the audience there indicated three quarters were using AI tools. A more empirical survey by McKinsey in the latter half of 2023, indicates that one-third of respondents say their organisations are using generative AI tools regularly in at least one business function. It would seem focus is shifting away from AI being a ‘tech thing’ as well. The same survey by McKinsey showed AI has risen from a topic owned by the technology department, to a focus of company leaders: nearly one-quarter of surveyed C-suite executives say they are personally using generative AI tools like Chat GTP for work.

The advanced machine learning algorithms that powers AI–enabled products has been decades in the making, your author cut their own technical teeth imagining and building early automated machine translation tools that used statistical machine learning algorithms and classical rule-based systems on human language…early runners to the large language models (LLM’s) we see today. It is worth noting Chat GPT was developed by OpenAI, an artificial intelligence research lab that was actually founded in December 2015 and the initial development of Chat GPT began in 2019. But since the tech as we know it today came off the starting block in November 2022, new iterations of generative AI technology have been released several times a month by a multitude of others…the floodgates are definitely open!

In the last 12 months we have seen Google, Amazon, Cohere, Microsoft and Meta all producing a slew of new LLM’s focused on domain specific goals such as human language translation, medical assistants, human prompting and reasoning, image and audio analysis, CRM tooling, and many, many more.

In March alone Chat GTP developers, released version 4 of the Chat GTP application. This offered a 40% improvement over the 3.5 release. 40% in four months. Incredible. In the same month Anthropic released a new type of AI engine called Claude trained using a method called constitutional AI, which aims to reduce the amount and likelihood of outputs that would be considered harmful. March also saw Google release its version of Chat GTP called Bard and Microsoft integrates Chat GTP into it’s Office 360 suite.

Our understanding of the potential threats and issues of AI is maturing too, and we are starting to have sensible conversations as a society about what we need and want AI to do, the potential impacts as we see them today and what measures we might want to consider to keep the human in the loop for as long as possible going into the future. To that end, the UK hosted the first global AI Safety Summit in November 2023, at the ancestral home of computing at Bletchley Park. It brought together leading AI nations, politicians, technology companies, researchers, and civil society groups to turbocharge action on the safe and responsible development of frontier AI around the world.

In total 28 countries from across the globe including Africa, the Middle East, and Asia, as well as the EU, agreed to the Bletchley Declaration on AI safety, which recognises the urgent need to understand and collectively manage potential risks through a new joint global effort to ensure AI is developed and deployed in a safe, responsible way for the benefit of the global community. Time will obviously be the judge of these noble intentions and it feels like we are at the very start of this conversation.

One thing is for sure, and in the indubitable words of 1970’s Canadian rock band, Bachman-Turner Overdrive - “you ain’t seen nothin’ yet”.

Happy Birthday Chat GTP!