đ¤ w/150: You Donât Have Time To Wait For AI Adoption In Your Business
This issue of Wiser! discusses the importance and implications of AI adoption in businesses, provides updates on AI developments, and offers advice on implementing AI in organisations.
w/Wiser! #150 - 3rd December 2023
Dear Wiserers!
So, here we are, issue 150! It feels like a major milestone to me, but one thatâs been over shadowed. Because itâs also been ChatGPTâs first official birthday!
Iâve been unable to find ChatGPTâs actual birthday, or even their day of conception, but like the King of England, ChatGPT has an official birthday, and it was this week. ÂĄFeliz cumpleaĂąos Chat!
Now onto something thatâs been occupying my mind this week. And Iâm not talking about Elon Muskâs ketamine fuelled breakdown on stage in New York this week, but something much more important for business leaders to ponder. The adoption of AI into the organisation.
Starting point: organisational change takes time. But there isnât much time. With the sort of efficiency gains we have seen in just the first year of ChatGPT, organisations that wait are going to fall behind. And quickly.
Iâm talking about adopting AI tools, processes and policies into any workflow, whether youâre a single person business to the largest of enterprises. If we can trim a week long process into a dayâs work, or boost the productivity of below average workers, these are profound changes to how work gets done.
So, the question becomes not whether to embrace AI but how and when. IMHO, the âwhenâ is easy, itâs now. Which leaves the question of âhow.â
The evidence suggests that employees are already embracing the âhowâ and using AI anyway. They just arenât telling their bosses about it! This recent (Oct 2023) survey from Salesforce found that over half of people using AI at work are doing so without their bosses approval. The survey also found that 69% of workers had no training in AI and 64% had passed off AI work as their own.
The point here is that regardless of the percentages, it should be extremely worrying to employers if any staff are using AI on their own volition and worse, passing it off as their own. There are a ton of risks around data, privacy, confidentiality, copyright et al that would send any risk officer or financial director into a tail spin!
However, the flip side is that by embracing AI across the organisation in a controlled way, you can minimise these risks AND create the foundation for organisational change and improvement.
Iâm going to generalise now, but broadly speaking my advice to decision makers on the âhowâ question can be broken down into five steps:
1- Create an AI council
Organisations need to develop their own guidance because there are no playbooks or blueprints out there. This is unchartered territory. And you donât want everyone going their own way because one favours ChatGPT, another likes Claude and the CTO wants Microsoft CoPilot. Create a cross-functional AI Council to represent all corners of the business. And include legal!
2- Prioritise Education and Training in AI
You have to get your knowledge workers on board and comfortable with the basics. In the same way as youâd expect staff to know their way around email, Word or Excel, all staff need to have a 101 understanding of how to access AI, the limitations and dangers of the tech, and how to prompt.
3- Develop AI Policies and Principles
The key here is to keep them short and easy to understand. There are two reasons for this. First, see the previous step. Understanding of AI across the organisation will be at a rudimentary level. The words need to be easy to understand by everyone! Second, the world is changing quickly and policies will change. You need to write them in pencil, not indelible ink
4- Start With The Low Hanging Fruit
Chalking up early wins is important to build confidence and support in this program of change. It is also a fast track learning curve with real feedback on what can and canât be expected. But thatâs not all, you also need a broader impact assessment across the whole of the knowledge worker base, paying particular attention to the impact on human roles.
5- Build an AI Road Map
This is harder than it sounds because the landscape is changing so quickly. However, what you can do is build a roadmap that lays out the priority projects youâre going to focus on and plan for the changes youâre expecting from the adoption of AI in each area.
Hereâs The Thing â I see many organisations talk about using AI but at best theyâre only playing around with the free versions of AI. Like using ChatGPT but not paying for ChatGPT Plus. The issue here is that their understanding of what AI can do for them does is limited to the capabilities of the free version.
Like any software platform with a free and paying subscription business model, those that pay get the goodies whilst those that take the free version get a no-frills, trimmed down version. So, all the good stuff, all the new features and capabilities, are given to the paying subscribers first. If these are your competitors, then they have an advantage over you already!
As we approach 2024, now is the time to be planning an preparing for the new year ahead. If youâve not done so already, now is absolutely the time to stop saying youâre gonna do it, and start embracing AI in your organisation.
w/MyWork
Californiaâs Bold Move to Lay Down The Law in AI
Capitalism beats Safetyism as OpenAI reinstate Sam Altman (YouTube)
A video update for Big Tech Little Tech.
Big Tech Little Tech is available on all major podcast platforms â đď¸Â Apple Podcasts | Spotify | Amazon Music | Anchor | YouTube
w/News
What Else Happened This Week?
Perplexity AI unveils âonlineâ LLMs that could dethrone Google Search
If youâre not using Perplexity already, then I urge you try it out as your Search engine. Iâve ditched conventional search engines in favour of Perplexity for all my searches because Perplexity gives you an answer, not a bunch of links for you to find the answer. You can run it within your browser, or download the app. Anyhow, the news is that theyâve just announced two new AI models that improve their already very impressive AI capability. Link
Inflectionâs Pi just got even better
Inflection AI is the startup behind my favourite AI, the conversational chatbot called Pi. Run by former DeepMind founder Mustyfa Sulleyman, Inflection has unveiled a new AI model that the company claims can outperform Googleâs Bard and Metaâs LLaMA chatbots. And is hot on the heels of OpenAIâs larger, flagship model GPT-4. More on this next week. Link
OpenAI announces Microsoft on the new all-male board
As I predicted in last weekâs Wiser!, the new structure of the OpenAI board includes Microsoft as a non-voting member. You can ignore the ânon-votingâ bit because, rest assured, as the largest shareholder (49%) theyâll have a big say in the running of the organisation. The other notable point is that the all new board lacks any female representation. At a time of great uncertainty about the risks posed by future AI, Iâd feel much happier if our future wasnât solely in the hands of a bunch of extremely wealthy, middle-aged tech men. Link / Read my blog: Profits Over Safety
Sports Illustrated caught using fake AI writers
The tech reporters at Futurism were first to notice that Sports Illustrated are the latest publication to be using fake people with AI-generated headshots to âwriteâ automated copy. In their words, âthe resulting product isnât very good.â Youâd think theyâd know better by now! Link
Researchers test AI models for intelligence, and find that theyâre not very smart!
Artificial intelligence programs still struggle with basic problem-solving skills that people excel at - is the finding from some new research from Metaâs top AI scientist. Which is bleeding obvious if you understand anything about large language models - they're word prediction machines and not "intelligent" at all. Link / Read my blog: AI is Remarkable Clever, and Shockingly Stupid
Anna Indiana is the world's first all-AI singer-songwriter
Anna Indiana is the name given to the world's first all-AI singer-songwriter. The lyrics, soundtrack, vocals and video of her/it singing are all AI generated using ChatGPT. Iâm not sure which pronoun is appropriate - Her/She/It/They - whatever. But you can watch the video of Anna singing her own song on YouTube and decide for yourself if itâs any good. To me it sounded like it belonged in a Disney movie. The bigger and more serious thing here is that this is just the beginning of AI generated music. Watch closely! Link / YouTube
Apple's Cook calls for AI Regulation, while Google's Schmidt warns of inadequacy in regs
Apple CEO Tim Cook has emphasises the need for regulations and guardrails to prevent misuse of artificial intelligence, calling for "rules of the road" in the industry. Despite optimism for AI's potential, Cook acknowledges the risks associated with AI technology, including deepfakes and the spread of AI-generated child abuse material. Cook suggests that AI regulation will likely be implemented within the next 12 to 18 months, with governments worldwide catching up on the need for regulation.
Meanwhile, former Google CEO Eric Schmidt goes the other way and warns that AI guardrails are insufficient to control AI capabilities. Schmidt proposes creating a global body similar to the IPCC to provide accurate information to policymakers and take urgent action.
Bottom line is that regulating AI is never going to keep up with the tech. Which partly explains why California is taking the approach of regulating the outcome for the user.
Schmidt Link / Cook Link / Read my blog: Californiaâs Bold Move to Lay Down The Law in AI
FREE TO WISER! READERS: If youâre new to generative AI and want to know the basics of how to get started, what you can do with AI and sample prompts to get you goingâŚdownload your copy of The Beginnerâs Guide To ChatGPT for FREE!
đ Show Your Support For Wiser!
Thank you for reading Wiser! If you got value and would like to support what Iâm doing, do this:
Forward this email: send it to anyone you know whoâs interested in the tech economy.
Make a donation: go to BuyMeACoffee to make a donation in the form of a virtual cup of coffee, they only cost âŹ2 each. And who doesnât love a coffee, right?
Check out my website: Youâll find links to everything I do at rickhuckstep.com.