The AI Boom is Over - Now the Real Work Can Begin

I’ve been seeing a lot of articles lately about the end of the “AI Boom” - which speaks to how quickly things change in the world of Artificial Intelligence. Less than a year ago (November, to be exact) OpenAI released ChatGPT 3.5. Shortly after, Vers. 4 was introduced, along with Google’s announcement of Bard. 

Since then, there has been no shortage of AI-related tools, chatbots, and plug-ins that either live on or assist the OpenAI platform. Google has also opened their GenAI LLM and Machine Learning tutorials to the public. I’ve personally worked with a variety of AI-content assistants, from GPT 4 and AIPRM, to Pencil AI and of course, Bard. There’s even a new Musk-backed AI model coming out (which I am somewhat skeptical about). 

Yet, the boom is already over. And with that, I say now we can focus on the real work of GenAI. We will not only begin to truly unlock the potential of things like AI-assisted content for marketers, creatives, and those who are looking to inspire others, but focus on making GenAI the best it can be. But how do we make AI work for everyone involved? There are three key things we should address to make our efforts shine:

Governance & Bias

Listen, I think we all know that AI has a HUGE bias issue and has from the beginning. There’s plenty of concerns still happening with even the latest updates in functionality. 

One recent example that has been making the viral rounds shows a woman of Asian descent, using an AI-supported image app to give her headshot a more ‘professional’ look. The image app turned her into more of a ‘white looking’ person, suggesting that white women are more ‘professional’ looking.

Another recent snafu included a recruiter and job placement specialist who had been utilizing GPT, and mentioned that he is Jamaican - following this, all of ChatGPT's messages to him started with “Big Up” and ended with “One Love.”

This is still a massive issue across many LLMs and chatbots. There needs to be a diversity and inclusion overhaul across all of artificial intelligence.

Accessibility to Everyone 

One of the initial things I enjoyed about AI was that basically anyone could use it. For GPT, you just needed an account. For Bard, you just had to get on a wait list. But with the growth of use from companies, the inevitable claws of capitalism are coming out. ChatGPT Enterprise is here, and for an undisclosed amount (likely based on your client’s spend and needs), you can have ‘extra special access’ to the OpenAI platform. Bard is free, still, for now on Workspace accounts, but it may only be a matter of time before Google locks down aspects of their GenAI behind costs.

For all the misgivings I have about Enterprise level GPT I’ll give OpenAI credit, they are thinking beyond the Enterprise user. The org recently shared their plans for ‘Teaching with AI’ which they bill as: “We’re releasing a guide for teachers using ChatGPT in their classroom—including suggested prompts, an explanation of how ChatGPT works and its limitations, the efficacy of AI detectors, and bias.”

Accountability

When it comes to the actual output of Large Language Models (LLMs), many are still black boxes of information (or in some cases, disinformation). Sourcing content and outputs are still something of a grievance to many, and rightfully so. Huge organizations are striking out (literally) against the use of AI content, including The WGA and SAG-AFTRA strikes for fair entertainment writer, actor and creative pay, along with a push against the use of AI-generated scripts and performers.

Google’s AI-supported search results, currently known as SGE or ’Search Generative Experience,’ just began citing their sources with links to information relevant to the AI-generated search result snippet. It’s the first step towards accountability in AI within search results.

There’s still a lot to do before AI can become the ‘perfect’ companion to those who will get the most use out of its unique aspects, whether that be for marketing, education, or creativity. It will happen fast, I predict, but until then AI will be…just as humans…perfectly imperfect.