Gen-AI Faces Impending Slowdown as Costs Surge, Hype Fades

Gen-AI Faces Impending Slowdown as Costs Surge, Hype Fades

Generative AI has garnered a lot of hype this year, but analysts see this fading in 2024 due to the high costs of running the technology coupled with regulation.

An analyst firm predicts that such issues could point to a slowdown in generative AI as the world faces the reality of running the technology. Yet businesses and CEOs of global companies are upbeat about the sector’s transformative power and are willing to invest in it.

Other key predictions by the company are centered around content warnings and identity fraud using AI technology.

A cold shower in 2024

In its roundup of 2024 forecasts for the tech sector, CCS Insight highlights what they think AI faces next year and beyond. Their key highlight is that despite making headlines during the year 2023, generative AI could “get a cold shower in 2024” due to the high costs of running the technology and the risks involved.

CCS Insight chief analyst Ben Wood told CNBC that generative AI was the buzz, with everyone talking about it as well as the key players in the sector like Amazon, Meta, Qualcomm, and Google.

“We are big advocates for AI; we think that it’s going to have a huge impact on the economy; we think it’s going to have big impacts on society at large; we think it’s a great opportunity for productivity,” said Wood.

“But we think the hype around generative AI in 2023 has just been so immense that we think it’s overhyped, and there’s lots of obstacles that need to get through to bring it to market.”

Since OpenAI launched its ChatGPT in November last year to immediate success, generative AI has gained prominence, with various tech companies coming up with their own models to rival ChatGPT.

Also read: OpenAI CEO Questions U.S. Crypto Approach, Backs Bitcoin Freedom

Unsustainable costs

But according to CCS Insight, the escalating costs of running the models will weigh on the sector in the short to mid-term. This comes as AI-focused firms are reportedly looking at ways to turn the generative AI hype into profitability.

Models such as ChatGPT, Google Bard, and Anthropic’s Claude use a lot of computing power to run “the mathematical models” that allow them to work out responses to users’ prompts.

“Just the cost of deploying and sustaining generative AI is immense,” Wood told CNBC.

“And it’s all very well for these massive companies to be doing it. But for many organizations and developers, it’s just going to become too expensive.”

According to the CNBC report, some companies in generative AI require high-powered chips—graphic processing units (GPUs)—to run their workloads.

With the rising costs, companies like Amazon, OpenAI, Meta, Alibaba, and Google are now working on producing their own specific chips to run these systems.

Other studies have also shown high energy and water consumption associated with generative AI, adding costs that are expected to continue rising as demand for the technology also escalates.

The regulatory conundrum

According to CCS Insight’s analysts, AI regulation faces obstacles, for instance in the European Union; the block has been leading in coming up with specific regulations for the sector, but these are likely to be redrawn several times in line with the fast pace of advancements in the sector.

“Legislation is not finalized until 2024, leaving industry to take the initial steps toward self-regulation,” said Wood.

The sector has created a lot of excitement on the back of its ability to generate content, prompting governments and individuals to call for its regulation as fears grew that the technology might pose threats to humans.

The technology has already been used to spread misinformation and deepfakes and there are calls for online platforms to watermark content that is generated by AI or by humans. Additionally, there are an estimated 49 news sites with content generated by AI software only.

CCS Insight has also predicted that police will make their first arrest of anyone who uses AI for impersonation or any form of identity theft.

“Image generation and voice synthesis foundation models can be customized to impersonate a target using data posted publicly on social media, enabling the creation of cost-effective and realistic deepfakes,” said CCS Insight in its prediction list.

“Potential impacts are wide-ranging, including in banking, insurance, and benefits,” added the firm.

Image credits: Shutterstock, CC images, Midjourney, Unsplash.