August marked a somewhat melancholic shift in people’s attitudes toward artificial intelligence. Yes, anti-AI spokespeople have been singing their concerns for a while. But August began to feel a bit different ever since OpenAI released GPT-5 on the 7th, with more folks voicing their concerns. In that wake, MIT released a study about AI initiatives’ ROI blunders, and there’s been snowballing talk of the AI bubble finally bursting.
On the flip side, Perplexity announced the next evolution of its Publishers’ Program on its mission to create a better, fairer internet for everyone.
Here’s your August AI roundup.
Subscribe to the ai marketer
Weekly updates on all things ai in marketing.
Google’s AI Mode Went Global
Before exploring some of the more tumultuous items of this round-up, here’s a ‘lighthearted’ one (depending on how you feel about the feature). Google’s AI Mode went global this August, adding hundreds more countries throughout the Americas, Asia-Pacific, Europe, the Middle East and Africa, where users can now access the feature. If you’ve been keen to try it out, check their list of supported countries and territories to see if it’s available where you are (or head to Google, where they make the option fairly obvious).

The international rollout was also accompanied by an announcement of new agentic capabilities for AI Mode, which Google says “can help you get things done more easily,” including tasks like booking reservations at a restaurant — personalized by party size, occasion, cuisine and more.
A New MIT Study Reports Most AI Investments Net Zero Measurable Returns
MIT says as much as 95% of enterprise AI pilots fail, netting no measurable return on investment.
The new report, titled “The GenAI Divide: State of AI in Business 2025”, examines the real-world outcomes of generative AI initiatives across businesses. Most notably, it highlights the gap between AI investment hype and actual returns, dubbing the phenomenon the “GenAI Divide”.
Here’s a small taste of what MIT researchers found:
- Despite $30–40 billion invested in enterprise GenAI, a staggering 95% of organizations are seeing zero return on investment.
- Only 5% of integrated AI pilots are generating significant value.
- Only 2 out of 8 major sectors — specifically tech and media — are showing meaningful structural change due to GenAI, while most others remain largely unaffected.
- 50% of GenAI budgets are directed toward sales and marketing, despite back-office automation often yielding higher ROI.
Depending on one’s disposition toward AI in general, encountering these numbers for the first time could be sobering or entirely unsurprising. However, some outlets have begun dissecting the research to uncover its flaws, including a narrow definition of success, ignoring metrics beyond ROI and data limitations.
Something else is interesting here, too: A survey from the report asked respondents to divvy up a hypothetical $100 budget across different business functions, 70% of which was funneled into sales and marketing functions, including AI-generated outbound emails and smart lead scoring (pg. 9). The report notes that sales and marketing dominate in this regard because “outcomes can be measured easily,” and, “Metrics such as demo volume or email response time align directly with board-level KPIs.”
So, even if returns are lacking or an AI bubble burst is on the way (more on this below), sales and marketing attitudes toward generative AI may remain unaffected, or at least little changed, since that’s where people prefer to push their AI budgets.
No matter if you’re a skeptic or supporter of AI, this new research provides lots to think about in terms of adoption, investment and the future of artificial intelligence as a whole.
Perplexity Continues Its Push To Please Publishers
This month, Perplexity launched a new standalone subscription called Comet Plus. This new tier provides users access to premium content from select trusted publishers and journalists, with 80% of subscription revenue paid directly to those entities, while 20% is reserved for Perplexity’s computing and operating costs. Specifically, revenue is allocated to Perplexity’s partners based on a few different types of internet traffic, including:
- Human visits.
- Search citations.
- AI agent actions.
“The revenue allocation recognizes the reality that users now choose how they want to consume information: browsing manually, asking for AI-generated answers, or deploying agents for complex tasks,” wrote the Perplexity Team. “Publishers deserve compensation that matches this new reality.”
The platform is on a mission to create a better and fairer internet, which, in its current state, they say is filled with “low-quality clickbait, high-friction experiences, and endless blue links to human slop and no answers.”
This feels like a great next step and evolution from Preplexity’s previous Publishers’ Program, which shared revenue from ad placements that appeared alongside responses citing publishers’ content.
At the end of their announcement, they allude that Comet will eventually become free for all Perplexity users. If you can’t stomach another subscription, you’ll still get to experience Comet in due time.
Is the AI Bubble About To Burst? Many Think So Following GPT-5 Release
Dialogue about a bursting AI bubble has been brewing for a long time. But, ever since OpenAI unveiled GPT-5 on August 7th, 2025, that simmering pot seems to be getting closer to boiling over.
The new model for ChatGPT was met with mixed impressions, with many users saying it performed worse than previous versions across several of its capabilities. This flop followed early promises about GPT-5’s alleged era-defining upgrades, which are more or less viewed as just another name change by now. But it triggered an uproar of voices about the AI bubble, with media outlets and thought leaders from all over discussing their takes on when it will ultimately burst. Meta — infamous for its AI spending this year — is even purportedly slowing down its activities as it restrategizes for the nth time in 2025.
Whether or not we’re close to another dot-com-type detonation with AI is unclear. Actionable advice for now? Continue doing what works best for your business, stay apprised of updates and take action accordingly.
Have We Finally Caught Up to Innovation?
As always, it’s impossible to know where all of this is headed for certain. But August events have negatively tinged AI talk perhaps more than ever before. More people are skeptical, big brands are allegedly slowing down their AI efforts and, to top it off, MIT says enterprise AI investment seldom brings measurable returns.
The sprint toward artificial general intelligence (AGI) is inevitable, no matter how we feel about it. But perhaps August rekindles an age-old reminder that taking it “slow and steady” is almost always better than rushing, only to burn out and come crashing down.