The Big Bet On Generative Tech & Its Impact on Humanity
How generative AI is creating a blooming application landscape that will radically change the world
What is Generative AI and Generative Technology?
Generative AI is a field of AI where the AI is generating something new rather than analyzing something that already exists, such as images, audio, videos, text, or code.
Generative Tech, as NfX explains, is the broader view of what generative AI enables, including all the applications that the AI models at the base layer of the technology stack make possible.
In other words, generative tech is about what we can do with generative AI.
What are examples of Generative Tech?
The most common use-case for generative tech so far has been generative AI image creation through text-to-image generators like Open-AI’s Dall-E and Stable Diffusion (in fact, the images used in this article were generated by Midjourney, another AI image generator).
These models have enabled creators to accelerate their creative processes and have been a major driver of the growth in the last 6 months of this industry.
The opening up of the infrastructure (model) layer that companies like these have done has led to the proliferation of generative use-cases such as:
Marketing - such as helping content teams write better content, faster
Software development - helping development teams write code faster
Research - helping researchers find information faster
Sequoia has been at the forefront of defining and investing in this new market and recently released a work-in-progress landscape map:
Image Credit: Sonya Huang - Partner at Sequoia
How can Generative Tech impact humanity?
The primary unlock for generative technology is actually to help us as humans be more productive, creative, and effective. While some generative AI use-cases might replace the need for humans in certain situations, much of it will actually be used by us humans to enhance ourselves.
Sequoia Capital claims generative AI can make creative and knowledge workers at least 10% more efficient and/or creative, increasing their production capabilities. If this is true, then it should mean that trillions of dollars of value can be unlocked.
“The real power of generative AI comes from how widely accessible it is”, states Maya Ackerman, PhD, CEO & Co-Founder at WaveAI, who also teaches AI & ML at Santa Clara University. “Just as the previous wave of AI - predictive AI - was integrated into virtually all industries, the same will happen over the next few years with generative AI methods. Everywhere people create, they will soon be creating together with generative AI”
Professor Ackerman isn’t alone in believing the huge driver of value comes from the democratization of this technology. Matt White, CEO & Cofounder of Berkeley Synthetic, which develops AI R&D for what they call ‘Synthetic Reality’ states, “generative AI will fuel the creator economy and has the potential to bring hundreds of millions of people out of poverty by reducing the barriers to entry for content creation and monetization.”
While the distribution of these technologies may technically be open globally, that doesn’t necessarily mean adoption will happen evenly. As we see with any technology, there’s a question of whether the enhancements in creativity and increases in production will be felt equally across populations or if Big Tech and already well-off people in big economies across the world will be the main benefactors.
But, if the saying “a rising tide floats all boats” is true, then the whole world should feel the benefits, regardless of how distributed the access to generative technologies ends up being over time.
What use-cases exist for impactful Generative Tech?
The opportunities to apply generative tech are far-reaching across industries and use-cases. And we’re still SUPER early. So early, in fact, that this ‘industry’ has just recently been categorized and recognized by analysts, investors and other market participants.
That said, everyone’s still trying to figure out what it will be useful for and what industries it will disrupt.
One thing is clear, however, and that’s that it will have a HUGE impact on our lives and our economy, acting as an accelerator to scientific and technological progress.
Here’s some examples of applications that could be developed:
Learning - Imagine anyone with internet access being able to develop their own custom curriculum based on what they want to learn and the AI spits out the syllabus and an AI teacher walks them through it.
Investment advice - Could we take robo advisors to a whole new level with hyper-customized, professional-grade financial advice based on your current situation?
Programming - Code building AI is already here, but it will continue to improve. You can imagine it will get to a point where non-technical people could easily create a software application without knowing how to code. This could massively unlock innovation in underserved parts of the world, reducing the barriers to developing technology.
Drug discovery - Generative AI is already proving useful in the drug discovery process, making the discovery and commercialization of new therapeutics faster and more efficient. If this gets translated into lower drug costs, it could lead to saving more lives of at-risk populations that might not have been otherwise able to afford it.
What are the risks of Generative Technologies?
This budding new sector doesn’t come without its risks. And the risks here can be pretty severe.
The three main issues come down to:
Bias - The underlying AI models may have biases based on the data used to train them
Usage by bad actors - People can use these tools for unethical purposes, such as how DeepFakes have been used to impersonate people and spread misinformation
Unintended consequences - For example, the issues around copyright infringement where artists are arguing the artistic works of image generation platforms are indiscriminately using their copyrighted images as inputs in the AI models
Eliminating Bias in AI
There can be all types of bias, including racial and gender, that manifest in the applications if the data used to train the systems are non-inclusive.
Since the underlying models are generally trained via public data scraped on the internet, it is very susceptible to understanding the world in a biased way, since the world is inherently biased in different ways.
There are ways to generate synthetic data that are being explored where we can build more inclusive and diverse data sets to try to eliminate biases, but we are still early in solving these problems.
Countering Bad Actors
There’s always going to be bad people doing bad things with any technology. Thankfully, tech can be used to try to stop them too.
Cybersecurity and other tech companies are already developing ways to spot DeepFakes - that is, using AI to develop hyper-realistic media (photo, video, audio) that impersonates a person.
We can imagine that in the future, advances in DeepFake prevention might look something like a browser plugin that detects the media you’re consuming for authenticity.
DeepFakes are just one example of how this technology can be abused, but there will be many more ways to come. Let’s just hope the good outweighs the bad.
Mitigating Unintended Consequences
These technologies can have major risks and negative externalities that we’re already seeing take place in the generative AI space.
Professor Ackerman recommends that “developers of generative AI systems should consider how their systems will be used, and make substantial, active efforts to foresee and mitigate misuse.”
While building fast with iterative design cycles is the norm for the tech industry, for some technologies like AI, perhaps it’s better to take a slower, more intentional approach because of the risks.
“Another important dimension is collaboration - developers and end users should work together to figure out how to best serve this powerful new technology to humanity”, she says.
Conclusion
Generative Tech is here but it’s in the first inning. We’ll see major advances to human productivity and many different applications that will benefit humanity and our planet.
While there are many risks, it’s critical that we advocate for the ethical development and use of this technology and invest in preventative measures and technologies to mitigate the risks. This includes building systems, especially at the infrastructure layer, intentionally and with collaboration across various stakeholders.
Lastly, if we ensure accessibility to this technology remains high and remove barriers and friction involved in its adoption, we can see the benefits of innovation and productivity by using it to reach across populations.