This past week the tech industry made its annual pilgrimage to the city of Toronto to discuss all things innovation at Collision 2023. The traveling conference (organized by parent company WebSummit) is designed to unite startups, investors, and industry giants. In previous years topics like crypto, NFTs or the blockchain were the bells of the ball. But this year there was only one topic on the tip of everyone’s tongues: artificial intelligence.
After three jam-packed days of speakers, panel discussions, tradeshow floors, and networking conversations (seriously, I forgot how exhausting conferences could be), a few key AI themes emerged. Some of these takeaways were expected, a fair amount were exciting, and a couple were slightly alarming. Let’s dive in!
Since ChatGPT broke onto the scene in November of 2022, generative AI and large language models (LLMs) have been on every business’s implementation hit list, and nowhere was this clearer than at Collision. From startups using OpenAI to provide services such as Chatbots and Virtual Assistants, to AI-Powered Search, and beyond, everyone is looking at how LLMs can solve their business problems.
It also became apparent that it’s not only startups have the desire to implement generative AI. Healthcare companies are looking to reshape how they diagnose patients. Agriculture organizations are looking to provide personalized advisories to farmers with crop-related questions. Every industry is exploring how learning models can help them innovate, and deliver, more efficiently.
Data is the New Oil
In the words of Jeff Weiner, CEO of LinkedIn, “Data really powers everything that we do.” It’s not outlandish to say that the collection, organization, cleaning, and interpretation of data has become the backbone of every modern business. As such, ensuring that data is private and secure was top-of-mind for attendees. Panels tackled topics such as encryption, walled gardens, and confidential computing technologies.
Once data is absolutely secure, companies are now looking to AI to make the most of it – whether that’s finding patterns, creating predictions, or optimizing workflows. Another problem for AI to solve? Data collaboration. How can companies bring together data stored across a variety of diverse systems to create a unified view and tell a single story?
Putting Customers and Employees First
It’s not to say that a focus on customer experience and employee experience isn’t exciting, it’s just that customer-centricity has become increasingly expected from contemporary businesses. It’s clear that AI is now at the forefront when it comes to serving customers, stakeholders, and employees. Solutions spanned across HR, accounts payable, customer service, e-commerce personalization (shameless self-plug) and so much more.
Leveling The Playing Field
A particularly thrilling topic for speakers and businesses was how AI can democratize access to the skills and talents needed to achieve goals – both for businesses and for each of us as individuals.
On the first day of the conference, Thomas Dohmke, CEO of GitHub, walked through the findings of a newly published study that analyzed how GitHub’s copilot AI software has been reshaping the industry. In addition, the report looks into how AI can enhance productivity and assist developers in developing countries. Furthermore, it says that AI can be beneficial by providing those in developing countries with the same tools used in developed nations.
Unity also took the stage at Collision to unveil Unity Muse AI (now in Beta). The goal of this new AI software? Enabling creators, no matter the skill level, to develop games and 3D experiences with nothing more than text-based prompts. These announcements and insights are in addition to the plethora of articles, how-to’s, and plug-ins for ChatGPT that empower anyone to create and organize written content.
Advances in AI aren’t just leveling the playing field for individuals. As organizations are facing a variety of constraints – from supply chains to limited budgets – they’re also looking to AI to help them compete with the big dogs.
One particular insight that stood out to me was from Kaz Nejatian, Shopify’s COO and VP of Product. While discussing Shopify's plans to implement AI for their small business partners, he reflected on his mother's one-woman e-commerce store. Kaz wanted to give her access to the same levels of SEO optimization, product description generation, and continuous testing and experimentation that large businesses with dedicated teams are able to oversee.
Phygital - maybe my least favorite portmanteau ever - refers to the marriage between physical and digital experiences.
The newly announced Apple Vision Pro was designed to “seamlessly blend digital content with your physical space”. Coach also recently unveiled an in-store augmented reality shopping experience. It’s clear that brands are buying in on blurring the line between the two. In doing so, they’re also striving towards an ‘always-on’ approach to building relationships with customers.
Some of the most exciting insights at Collision came from Geoffrey Hinton, commonly known as the ‘Godfather of AI’. (And more recently known for his departure from Google amid concerns about technology’s threat to humanity. But we’ll get into that later.)
When asked about where AI was headed, he identified multimodal large language models (MLLMs) as the next step in deep learning's evolution. Regular LLMs work with only one input (for instance, text-based prompts or image recognition). MLLMs will instead use neural networks to synthesize multiple formats of input as they learn. This could include visual, gestural, spatial, audio, and even speech recognition signals!
Biases and Discrimination
AI can perpetuate the biases of the system that it’s trained on, or worse, the biases of the humans that have designed it. This can lead to blatant discrimination. For example, facial recognition software continues to incorrectly identify darker-skinned individuals 5 to 10 times more than light-skinned participants.
But how can we combat the potential for AI discrimination? Test-running algorithms in real-life, and continuing to use human intelligence with human-in-the-loop processes. These first steps will allow for continuous input and monitoring from both sides.
Regulation and Risk
Regulations have always lagged behind evolving technology. As technology continues to advance more and more quickly, this lag time only becomes more pronounced. Lawmakers are beginning to propose regulations, such as the EU AI Act and the California Automated decision tools bill.
However, that’s not stopping industry from continuing to charge forward with using AI applications. As legislation lags behind, it’s important for all stakeholders to balance AI-led growth with responsibility. We must all consider the broader impact of these technological advances.
Geoffrey Hinton was most critical of AI’s potential for harm. This included warring echo chambers, perpetuating fake news, and even battle robots (which were not words I was expecting to hear on the Collision main stage.)
Is AI coming for our jobs?
While a Goldman Sachs prediction recently exclaimed that over 300 million jobs would be lost or degraded by AI, the mood at Collision wasn’t quite that grim. In fact, in a presentation by Cassie Kozyrkov, Chief Decision Scientist at Google, there was excitement about the prospect of AI. Cassie believes that we can automate what she called the “thunking” work, so that we can focus more on the “thinking” work.
Yes, during any time of industrial upheaval, there is a shift in the jobs being done - the example used by Cassie was of a distant aunt who was a ‘computer’ before the advent of the screen that you’re reading this on.
As AI shifts entire industries forward, there’s the expectation that jobs will be lost, but that other jobs will be gained, the ability to focus on what really matters in a role will increase, and productivity will skyrocket. One presenter even referred to the period of disruption that we’re in as the “productivity revolution” (here’s hoping!)
So, Is AI Just Hype?
Looking at what happened with the crypto bubble after how prominently it was highlighted at last year’s Collision does make me a bit more bullish to hop on the AI bandwagon. However, I do think that in the future it will fundamentally impact almost every aspect of our lives. Though whether that’s five years, 15 years, or 25 years down the road, it’s hard to say.
One thing’s for certain: the potential benefits of AI are so far-reaching that we’ve only just begun to scratch the surface of the problems that it can help solve. And as we move forward, it’s up to us all to do so responsibly, with respect for the technologies we design and their real-world implications.
At Granify, we know the importance of responsibly building and deploying AI. That’s why the Granify Brain, our Machine Learning technology, collects no Personally Identifiable Information (PII) and doesn’t use third-party cookies while using real-time data to boost conversions. Request a consultation with our experts to uncover how your organization can leverage the full power of AI.
You May Also Like
These Related Stories