The The Transform Technology Summits begin October 13 with Low-Code / No Code: Enabling Enterprise Agility. Register now!
A recent New York Times article concludes that new AI-powered automation tools like Codex for software developers will not eliminate jobs but will simply be a welcome help in increasing the productivity of programmers. This is consistent with the argument we hear more and more that people and AI have different strengths and that there will be appropriate roles for each.
As stated in a Harvard Business Review article: “AI-powered machines are faster, more precise, and always rational, but they are not intuitive, emotional, or culturally sensitive. The belief is that “AI plus humans” is something of a centaur, bigger than either one operating alone.
This idea of humans and AI producing better results has become a tenant of faith in technology. Everyone’s talking about human beings being released to perform high-level functions, but no one seems to know exactly what those high-level functions are, how they translate into actual work and jobs, or how many people are needed to perform. execute them.
A corollary of this augmented workforce narrative is that not only will AI-enhanced work allow people to pursue a higher level of abstract thinking, but, some argue, it will also elevate all of society to a higher level. higher standard of living. It is certainly an optimistic view, and we can hope for it. However, it could also be a story steeped in magical thought, with the real end of the game being a fully automated job.
What does the evidence tell us?
Do not mistake yourself ; There is evidence to support the view that AI will help us work rather than take our jobs. For example, the AI lab DeepMind is designing new chess systems so that the two intelligences work in tandem with humans rather than opposing them.
And Kai-Fu Lee, the Oracle of AI, also adheres to that promise. In his new book, AI 2041: Ten visions for our future, he argues that repetitive tasks ranging from stacking shelves to processing data will be done by machines, freeing up workers for more creative tasks. Forrester Research also explained that deploying AI enables people to better use their creative skills.
But, of course, some people are more creative than others, which means that not everyone would benefit to the same degree from the work enhanced by AI. This in turn reinforces the fear that AI-powered automation, even in its increased work capacity, could widen already existing income disparities.
One problem with the AI augmented workforce promise is that it tells us that AI will only do the repetitive work that we don’t want to do. But not all AI jobs are routine or boring.
Look no further than the role of the architect of semiconductor chips. It is a highly sophisticated profession, an advanced application of electrical engineering in arguably one of the most complex industries. If there had ever been a job that could be considered immune to AI, this would have been a good fit. Yet recent advances by Google and Synopsys (including using reinforcement learning neural network software) have shown the ability to do in hours what often took months for a team of engineers.
An ever-loyal observer of technology has always argued that algorithms “will optimize and speed up the tedious parts of the design process so that designers can focus on the crucial calls that require higher-level decision-making.”
A step on the road to fuller automation
More than likely, the current perception of AI-augmented work is a reflection on the current state of technology and not a precise vision of the future when automation is much more advanced. We first saw the potential of neural networks ten years ago, for example, and it has taken several years for this technology to develop to the point where it has practical benefits for consumers and businesses. Fueled in part by the pandemic, AI technology is now widely implemented. Even massage therapists should take note, as a robot massager can now offer deep tissue massage. Yet these are still the early days for AI.
Legend: EMMA from AiTreat, a robot that uses artificial intelligence to deliver massages. Source: CNN
Advances in AI are being driven by improvements in both hardware and software. The hardware side is governed by Moore’s Law, the idea that semiconductors improve by about twice the number of transistors – producing roughly equivalent performance and energy efficiency gains – both years (and similarly reduce computational costs). This principle has been credited with all kinds of electronic advancements over the past decades. As stated in a recent IEEE Spectrum article: “The impact of Moore’s Law on modern life cannot be overstated. We can’t get on a plane, make a call, or even turn on our dishwashers without feeling the effects. Without it, we wouldn’t have found the Higgs boson or created the Internet. Or have a supercomputer in your purse or pocket.
There is reason to believe that Moore’s Law improvements in computing are coming to an end. But cutting edge engineering, ranging from “chiplets” to 3D chip packaging, promises to keep the gains coming, at least for a while. These and other improvements in semiconductor design have led a chipmaker to promise a 1000-fold performance improvement by 2025!
Equally impressive are the improvements expected from AI software. GPT-3, the third iteration of OpenAI’s Generative Pre-trained Transformer, is a neural network model composed of 175 billion parameters. The system has been shown to be able to generate consistent prose from a text prompt. This is what it was designed for, but it turns out that it can generate other forms of text as well, including computer code and can generate images as well. Additionally, while it is believed that AI will help people be more creative, it may already be capable of creativity on its own.
When it launched in May 2020, GPT-3 was the largest neural network ever introduced, and it remains one of the largest dense networks. neural networks, only surpassed by Wu Dao 2.0 in China. (With 1.75 trillion parameters, Wu Dao 2.0 is another GPT-like language model and possibly the most powerful neural network ever.)
Some expectation is that GPT-4 will also grow taller and contain up to a trillion parameters. However, OpenAI CEO Sam Altman said it won’t be bigger than GPT-3 but will be much more efficient thanks to improved data algorithms and fine tuning. Altman also hinted at a future GPT-5. The point is, neural networks have a long way to go in size and sophistication. We are indeed in the midst of an era of AI acceleration.
In the new book, The robot rule: how artificial intelligence will transform everythingAuthor Martin Ford notes that “almost all tech startups are now investing, to some degree, in AI, and companies large and small in other industries are starting to deploy the technology.” The pace of innovation will only accelerate as capital continues to flow into the development of AI. Obviously, everything we are currently seeing in the path of AI-powered automation, including the belief that AI will help us work rather than take our jobs, is just a step away. preliminary for all that is yet to come. As for what is to come, it remains the domain of speculative fiction.
In To burn: A novel about the real robotic revolution, a Yale-trained lawyer was among those affected when his firm replaced 80% of the legal staff with machine learning software. It could happen in the near future. The remaining 20% were indeed augmented by AI, but the 80% had to find other work. In his case, he ends up working as an online personal assistant for the wealthy. Currently, the start-up Yo Labs is working to realize a variation of this vision. The company initially offers a mix of human and AI services, starting with a living, breathing assistant that relies on data to tackle subscriber to-do lists. It will be revealing to see if these assistants will be like the secretaries of old, but using AI, or if they will be displaced cognitive workers.
The AI-based transition to a largely automated world will take time, perhaps a few decades. This will lead to many changes, some of which are very disruptive. Adjustments will not be easy. It is tempting to think that in the end it will enrich the quality of human life. After all, as Aristotle said: “When the looms weave on their own, the slavery of man will end. But embracing AI’s augmented work concept as it is currently articulated could make us forget the potential risks of job loss. Kate Crawford, an academic who specializes in the social and political implications of technology, believes that AI is the deepest story of our time and that “a lot of people sleepwalking in it.”
We all need to have a clear understanding of the growing potential for disruption and prepare as best we can, largely by learning the skills most likely to be needed in the age to come. Businesses must do their part by providing skills training, and retraining will increasingly need to be a near-continuous process as the pace of technological change accelerates. The government needs to develop public policies that orient market forces driving automation towards positive outcomes for all, while preparing for a growing social safety net that could include a universal basic income.
Gary Grossman is Senior VP of Technology Practice at Edelman and Global Lead of the Edelman AI Center of Excellence.
VentureBeat’s mission is to be a digital public place for technical decision-makers to learn about transformative technology and conduct transactions. Our site provides essential information on data technologies and strategies to guide you in managing your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the topics that interest you
- our newsletters
- Closed thought leader content and discounted access to our popular events, such as Transform 2021: Learn more
- networking features, and more
Become a member