Generative AI can only be a harbinger of AI singularity

Generative AI, which is based on Large Language Models (LLMs) and transformative neural networks, has certainly caused quite a stir. Contrary to hype cycles around new technologies like the metaverse, crypto and Web3, generative AI tools like Stable Diffusion and ChatGPT are poised to have huge, potentially revolutionary impacts. These tools are already disrupting multiple fields, including the movie industry – and are a potential game changer for business software.

All of this has led Ben Thompson to declare in his own strategy newsletter to declare generative AI advancements mark “a new era in technology”.

Still, broadly speaking, it’s still early days for AI. On a next Normal English podcast, Thompson said AI is “still in the first inning”. Rex Woodbury in his Digital native newsletter agreed: “We are still in the early days of AI applications and leaps and bounds are made every year.” A story from the New York Times mention that this has led to a new “AI arms race”. More companies are expected participate in this competition “in the coming weeks and months.”

A harbinger of AI singularity

Now that the generative AI era has been duly anointed, what could be the next leap or era and when might that happen? It would be comforting to think that we will all have plenty of time to adjust to the changes that come with generative AI. But just as a foreshock can predict a major earthquake, this new era could herald an even bigger event, the coming AI singularity.

AI singularity refers to two concepts: The first defines “singularity” as a point where AI surpasses human intelligence, leading to rapid and exponential advancements in technology. The second refers to the belief that technology will be able to improve itself at a faster rate, leading to a point where technological progress is so rapid that it exceeds human ability to understand or predict it.

The first concept sounds exciting and promising – from developing cures for previously incurable diseases to solving nuclear fusion leading to cheap and unlimited energy – while the latter raises terrifying Skynet-esque concerns.

Even Sam Altman – CEO of OpenAI and a leading proponent of generative AI and the developer of ChatGPT and DALL-E 2 – has expressed concern. He said recently that a worst-case scenario for AI “turns out the lights for all of us”. He added that it is “impossible to overstate the importance of AI safety and fine-tuning work.”

When will the singularity come?

Expert predictions for when singularity arrival varies significantly; the most aggressive creature it will be Very fast. There are others who say it will be reached sometime in the next century, if at all. The most quoted and one of the most credible is futurist Ray Kurzweil, currently a director of engineering at Google. He famously predicted the arrival of the singularity in 2045 in this 2005 book The singularity is near.

Deep learning expert François Chollet similarly notes that predictions of the singularity are always 30 to 35 years away.

Still, it seems more and more that Vernor Vinge’s prediction comes closest. He coined the term singularity in 1993 article with a striking statement: “We are on the brink of change similar to the emergence of human life on Earth.”

Translateda start-up for translations into Italian recently claimed that the singularity occurs when AI provides “a perfect translation”. According to CEO Marco Trumpetti: “Language is the most natural thing of man.” He adds that language translation “remains one of the most complex and most difficult problems for a machine to perform at the level of a human” and is therefore a good proxy test to determine the arrival of the singularity.

To assess this, the company uses Mate cat, an open-source computer-aided translation (CAT) tool. The company has been tracking improvements since 2011 using Time to Edit (TTE), a metric in the tool to calculate the time it takes for professional human editors to correct AI-generated translations compared to human translations.

Over the past 11 years, the company has seen strong linear performance improvements. Based on this, they estimate that the time needed for perfect machine language translation will be towards the end of this decade, at which point the singularity will have arrived.

How will we know when the singularity arrives?

Of course, TTE is only one measure and need not in itself indicate a groundbreaking moment. As described in a Popular Mechanics article“it’s extremely difficult to predict where the singularity will start.”

It may be hard to locate, at least at the time. It probably won’t be a single day where a certain stat is achieved. The impact of AI will continue to grow, with the inevitable peaks and troughs of progress. With every advancement in AI, the tasks it can accomplish will increase.

There are already many signs of this, including DeepMind’s AlphaFold, which predicts the folding pattern of virtually any protein and could lead to radical improvements in drug development.

And Meta recently unveiled “Cicero”, an AI system that beat humans in Diplomacy, a strategic war game. Unlike other games that AI has mastered, such as chess and Go, Diplomacy is collaborative and competitive at the same time. As reported by Gizmodo“to ‘win’ at diplomacy [Cicero]one must understand both the rules of the game efficiently [and] fundamentally understand human interactions, deceptions and cooperation.”

Whisper was born late last year to finally produce fast and reliable voice-to-text transcripts of conversations. According to The New Yorkers, decades of work have led to this. Based on open-source code from OpenAI, it’s free, works on a laptop and (according to the reviewer) is way better than anything made before.

What can be the impact?

Identifying the advent of singularity is complicated because there is no universally accepted definition of what intelligence means. This makes it problematic to know exactly when AI will become more intelligent than humans. What can be said is that the capabilities of AI continue to evolve and at a breakneck pace.

Even if it hasn’t reached the singularity yet – and perhaps never will – the list of AI achievements continues to expand. The consequences of this, both good and bad, will also increase. One day, possibly in a few decades, there could be a ChatGPT-esque moment where the world will shake again, even more so than with Generative AI. With the “great” it will be assumed that the singularity has arrived.

It’s good to keep in mind what computer scientist and University of Washington professor Pedro Domingos said in his book The main algorithm: “Man is not a dying twig on the tree of life. Rather, we’re about to start branching out. In the same way that culture evolved with larger brains, we will evolve with our creations. We always have: Humans would be physically different if we hadn’t invented fire or spears. We are Homo technicians as much as Homo sapiens.”

Gary Grossman is SVP of technology practice at Nobleman and global leader of the Edelman AI Center of Excellence.