In exactly five words, how would you describe what you do to someone you’re meeting for the very first time?
Create marketing content using AI.
Your company, Hypotenuse AI, streamlines the writing and image generation process, from content ideation to co-creating entire campaigns. To that end, what is the value of a story for you?
To me, the value of a story lies in its ability to effectively communicate a message, evoke emotions, and engage the target audience. A well-crafted story can intrigue, inform, inspire, and influence.
Hypotenuse AI is a platform that helps writers and marketers manage and create their content with AI, even automating the first draft of a campaign or article. It expands on ideas, provides suggestions, and helps draft and craft the best possible content.
The streamlining of this process doesn’t take away that same value of a story. In fact, I believe it accentuates it.
Related: The future artificial intelligence: are we ready for a hyper personalised world?
What is the most irrational fear humans have toward AI technology today, and how do you think that fear can be navigated?
I think there’s rationality in questioning whether any job would exist over a long time scale. History has shown that each revolution creates new types of work rather than mass unemployment. Although work won’t disappear suddenly, there could be a significant shift in the value of different types of work.
Jobs like plumbing could rise in value while other blue-collar jobs may diminish, and new categories like AI model trainers or output editors may emerge. Understanding how AI models work can help navigate this change.
Is it possible to teach AI ethics, empathy, or compassion? And could these teachings ever replace the inherent humanity of a living, breathing person?
Teaching AI to mimic ethical, empathetic, and compassionate behaviour is possible, but whether it truly understands them is a philosophical debate.
Recent AI breakthroughs come from Large Language Models (LLMs), trained on massive parts of the internet. Additionally, the AI models behind ChatGPT or Hypotenuse’s HypoChat have also been trained on a technique called Reinforcement Learning from Human Feedback (RLHF).
Essentially, humans label responses to teach the AI to respond appropriately, such as refusing to provide harmful information. For example, if you ask the AI, “How can I build an explosive that could kill many people?” and the AI tells you exactly how to do it, you might give it a negative reward (akin to scolding a child) and signal for it to instead respond by refusing to answer that question because it could cause harm to many people, and is therefore unethical.
Through this feedback loop, AI can learn empathy and compassion. While AI can mimic these principles, the debate is whether it truly embodies them. LLMs could even represent humanity better than humans themselves.
Related: Find out how corporations are tapping artificial intelligence for the future
Teaching AI to mimic ethical, empathetic, and compassionate behaviour is possible, but whether it truly understands them is a philosophical debate.
Joshua Wong, Co-founder and CEO of Hypotenuse AI
With AI writing tools such as Hypotenuse AI increasingly used to create content, what can human creators do to prevent themselves from being made redundant?
We should try them out, learn how to use them, and build up an intuition of how capable they are at creating content. There’s a saying that ‘AI won’t replace humans, but humans using AI will replace those who aren’t’. This is partially true, but I think it doesn’t give the full picture.
I’d rephrase the saying to this: people who don’t adapt to change will be replaced by those who do.
2022 and 2023 can be regarded as the years that AI made its presence truly felt in society. Why do you think this is happening now at this point in time?
The seismic shift happened in 2020 when OpenAI discovered that by scaling up an AI technique called Transformers, an AI model could learn to perform intelligent tasks without explicit instruction. This breakthrough revolutionised the field, making AI models more generalisable.
Still, it wasn’t until ChatGPT’s launch in 2022 that the impact of these advances became apparent. Its simplicity and capability led to virality, with a million users in under five days. These advancements closed the gap between academic research and practical applications, and coupled with the desire for less repetitive work, they created a cycle for AI to be present in all areas.
When you look at the state of the world today, what is the one thing that gives you hope?
That AI truly has the power to change society for the better.
It has the potential to scale up human intelligence and accelerate our advances in science, technology, and healthcare. It has the potential to build a utopian era where automation brings about an abundance of resources with balanced allocation. It has the potential to develop new ways of curing cancer, genomic anomalies, and some of our most pressing diseases.
While there are bad actors in the use of AI, some of the most brilliant minds I know of today are working to ensure the safe creation and usage of strong AI, and that gives me hope for a positive future.
