From ChatGPT 3.5 to AGI: How big a leap?

It’s an interesting time in tech, especially with the ChatGPT craze that has put Microsoft front and center again and has even buoyed the hopes of companies like Micron that have seen a 13-year low in memory and storage but know that generative AI will inevitably fire up demand for memory chips.

On the one hand, companies want to try out ChatGPT 3.5 for cost savings with such tasks as content creation and data analysis. On the other hand, hundreds of tech leaders and scientists have recently argued against large experiments with AI that are more powerful than today’s ChatGPT and can pose profound risks to society.

There’s a great deal of middle ground between those two viewpoints, and it’s worth tracking what’s being said.

A survey by a workforce management software company called Workyard looked at 1,000 small to mid-sized digital companies and found 40% are already using ChatGPT 3.5 for automation. It isn’t clear how much the companies are relying on the tool, but Workyard concluded that it could result in cost savings.

For example, Workyard said ChatGPT 3.5 for automating social media management could possibly cut the cost of recruiting and overseeing social media managers by up to 90%. A simple online post for a product when sent to ChatGPT 3.5 under the supervision of strategists could save up to $200 a post, according to Workyard.

Using the tool in email outreach can save up to 24 cents per email based on the average cost of sending an outreach email that can reach 20 cents to 30 cents, including labor, email marketing software, data acquisition and more.

Also, for blog posts, content creation savings could be from $90 to $300 on average per post, compared to the normal cost of $100 to $500 to create social media captions, blog post headers and product descriptions.

One of the biggest areas of focus for chatbots has been for customer service. Workyard said using AI software and downsizing customer service teams can reduce employee expenses by $15,000 a month. ChapGPT- based AI tools cost $500 a month, compared to $3,000 per worker in a 10-person team.

Workyard also said a company could save nearly $4,000 per month by automating simple data analysis operations, saving nearly $4,000 a month over hiring an entry-level data analysts.

With ChatGPT 4, Workyard predicted cost savings up 30% to 60% for jobs involving video content creation, online tutorials, advanced translation and even programming.

Workyard  also addressed the biggest worries with AI—job displacement, biased algorithms and privacy.

“A balanced approach is essential, ensuring AI complements humans skills rather than replacing them and establishing safeguards against misuse,” Workyard CEO Nic De Bonis said in a statement. “Despite GPT 4’s advancements, AI models are not perfect and human intervention remains crucial for high-quality output in areas where nuances matter.”

De Bonis touched lightly on the downsides of generative AI, but others are more strident.

A group of more than 1,500 technology leaders and scientists have recently been adding their signatures to an open online letter calling on all AI labs to immediately pause for at least six months the training of AI systems more powerful than GPT-4.

Future of Life Institute posted the open letter and its signatories include Apple co-founder Steve Wozniak and Tesla CEO Elon Musk.

“Contemporary AI systems are now becoming human-competitive at general tasks,” the letter reads. “Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization? Such decision must not be delegated to unelected tech leaders.”

The pause called for in the open letter “should be public and verifiable and include all key actors. If such a pause cannot be enacted quickly, governments should step in an institute a moratorium.”

There is urgency, according to experts. OpenAI plans to release another version, GPT-4.5 in about six months with GPT-5 in a year.  Some expect GPT-5 to reach the level of AGI (Artificial General Intelligence), the point where AI becomes sentient, or thinking for itself. From there, it could grow quickly.

“Once you get to AGI, it’s like game over for human beings…It escalates completely out of control once you get to AGI. So that’s the big fear. At that point, humans have no control. It’s just out of our hands,” said Avivah Litan, a vice president and distinguished analyst at Gartner Research, in comments to Computerworld.

Microsoft had no comment on the open letter. The company backs OpenAI, the creator of GPT-4, with an investment of more than $10 billion.

Between the Workyard survey showing today’s sizeable interest in using ChatGPT 3.5 and a world with sentient beings, in GPT-5, there might be a year or more for companies and researchers to react.  Or not.

RELATED: Micron revenues tank but CEO Mehrotra sees promise with AI and its memory demands