Day by day, new AI integrations and advancements are popping up all across our work and personal lives. Here’s our Technical Director Tim’s take on the latest developments and disruptions we’re seeing as this technology continues to transform at a rapid pace.

No matter where you look, companies are touting their AI integrations. Your Spotify playlists are now powered by AI and Spotify Podcasts will soon be available in any language in the host’s actual voices, Microsoft invested USD$13 billion into Open AI and now wants you to create all your emails and documents with Microsoft Copilot, AI images are generating massive amounts of engagement from naïve boomers on Facebook, and travel companies are not being left behind either.

We are just at the tail end of a wide range of product announcements from the biggest companies in technology, with new tools and features coming out across the board.

OpenAI

Not one to hide in the shadows, OpenAI (the creator of ChatGPT) started the latest round of announcements a couple of days before the Google I/O event, hoping to steal some of their thunder. While there were some hopeful predictions of the next generation of ChatGPT or a GPT-powered search, the creators OpenAI released an improved version of the existing model, called ChatGPT-4o.

The new version is faster, can take conversational voice inputs, produces better images, but the biggest announcement is that a future update to 4o will include a new voice assistant. The assistant can be interacted with via audio or video, and is able to contextually understand where you are and what you’re doing. The pre-launch video examples show the assistant describing the user and the room they’re in while making assumptions about why they’re there and that they’re doing. The personality of the assistant has been described as far too flirty, and had already landed OpenAI into some hot water for replicating actress Scarlett Johanssen’s voice (in a nod to her role as a voice assistant in the 2013 Spike Jonez film Her) after reaching out to her and being told that she didn’t want to licence her voice for the project.

The ChatGPT-4o assistant is the first product with realistic, natural sounding voice interactions, and is sure to lead the way for a new generation of digital assistants/agents with unsettlingly natural personalities and conversational styles.

Limited access to the new model is available for free users, but paid users get more interactions.

Google

Google’s annual developer conference Google I/O showed their ongoing commitment to add AI to EVERYTHING. The Chrome browser is getting an inbuilt AI assistant, Android phones are getting AI spam detection, custom chat bots are coming to Gemini, Google Photos can be searched by AI, Generative video creation is coming, and Google Astra will either become the best virtual assistant in the world or never be mentioned again in true Google fashion.

The BIG announcement that took more than a few people by surprise is that the Google Generative AI search is not only leaving the experimental Google Labs for wide release, but has been significantly reimagined so that the entire results page can now be generated to answer a search. Likely responding to the pressure that newcomer Perplexity.ai is providing, Google are moving away from simple summaries of search content and reimagining a world where “Google does the Googling for you”.  The new Google will make suggestions, categorise results, and summarise information based on your questions. This new version of search will be rolled out starting in the US, with plans to reach over a billion users by the end of the year.

One of the strangest things to come out of the new search is a “Web” tab on the results page, where users can see JUST website results without the usual Google tiles, snippets, knowledge graph information etc. It feels like a return to a pure, original Google experience.

The response to these changes is near-universal concern. A report from the Wall Street Journal suggests an expected 20-40% drop off in organic web traffic as AI responses lead to more Zero-click searches (a search where no links are clicked on the results page) than ever before.

Microsoft

The most recent developer conference came in the form of Microsoft Build, where Microsoft continued to build on their investment in OpenAI and their desire to incorporate their AI Copilot tool into everything.

As a work productivity company at its core, it’s unsurprising that they’re pushing quickly into AI “Agents” that will be able to work like virtual employees and automate repetitive or time consuming tasks. They have the examples of Copilot being able to do data entry, answer HR tasks, and reduce administration time. A native connection to SharePoint and OneDrive should make training the agents on your companies information incredibly easy and make the agents very customisable for your business needs.

A big focus was also on hardware, where a new type of Copilot+ Laptops are being released with most major companies including HP, Asus, HP and Dell. These new laptops are meant to compete with Apple’s lineup and have AI built-in from the ground up, including an interesting feature called “Recall” where AI will have a searchable memory of everything you have seen and done on your computer. Incredibly useful if you have a habit of forgetting where you save files or what you had open before a shutdown, but also raises some interesting privacy questions. How do you feel about your computer remembering everything that you do on it?

Curious to learn more? Check out our AI Hub for the latest news and updates you need to know.