August AI Roundup & Rundown: Google’s AI Mode Goes World, AI Funding Nets No Return and a Bubble Sure To Burst?

August marked a considerably melancholic shift in folks’s attitudes towards synthetic intelligence. Sure, anti-AI spokespeople have been singing their issues for some time. However August started to really feel a bit totally different ever since OpenAI launched GPT-5 on the seventh, with extra people voicing their issues. In that wake, MIT launched a research about AI initiatives’ ROI blunders, and there’s been snowballing discuss of the AI bubble lastly bursting.

On the flip aspect, Perplexity introduced the following evolution of its Publishers’ Program on its mission to create a greater, fairer web for everybody.

Right here’s your August AI roundup.

Google’s AI Mode Went World

Earlier than exploring a few of the extra tumultuous objects of this round-up, right here’s a ‘lighthearted’ one (relying on how you are feeling concerning the characteristic). Google’s AI Mode went international this August, including lots of extra nations all through the Americas, Asia-Pacific, Europe, the Center East and Africa, the place customers can now entry the characteristic. In the event you’ve been eager to attempt it out, test their checklist of supported nations and territories to see if it’s accessible the place you might be (or head to Google, the place they make the choice pretty apparent).

The worldwide rollout was additionally accompanied by an announcement of latest agentic capabilities for AI Mode, which Google says “will help you get issues completed extra simply,” together with duties like reserving reservations at a restaurant — personalised by get together dimension, event, delicacies and extra.

A New MIT Research Reviews Most AI Investments Internet Zero Measurable Returns

MIT says as a lot as 95% of enterprise AI pilots fail, netting no measurable return on funding. 

The brand new report, titled “The GenAI Divide: State of AI in Enterprise 2025”, examines the real-world outcomes of generative AI initiatives throughout companies. Most notably, it highlights the hole between AI funding hype and precise returns, dubbing the phenomenon the “GenAI Divide”.

Right here’s a small style of what MIT researchers discovered:

• Regardless of $30–40 billion invested in enterprise GenAI, a staggering 95% of organizations are seeing zero return on funding.
• Solely 5% of built-in AI pilots are producing important worth.
• Solely 2 out of 8 main sectors — particularly tech and media — are exhibiting significant structural change because of GenAI, whereas most others stay largely unaffected.
• 50% of GenAI budgets are directed towards gross sales and advertising and marketing, regardless of back-office automation typically yielding increased ROI.

Relying on one’s disposition towards AI typically, encountering these numbers for the primary time may very well be sobering or fully unsurprising. Nonetheless, some shops have begun dissecting the analysis to uncover its flaws, together with a slender definition of success, ignoring metrics past ROI and information limitations.

One thing else is attention-grabbing right here, too: A survey from the report requested respondents to divvy up a hypothetical $100 price range throughout totally different enterprise capabilities, 70% of which was funneled into gross sales and advertising and marketing capabilities, together with AI-generated outbound emails and sensible lead scoring (pg. 9). The report notes that gross sales and advertising and marketing dominate on this regard as a result of “outcomes might be measured simply,” and, “Metrics akin to demo quantity or e mail response time align instantly with board-level KPIs.”

So, even when returns are missing or an AI bubble burst is on the way in which (extra on this under), gross sales and advertising and marketing attitudes towards generative AI could stay unaffected, or a minimum of little modified, since that’s the place folks favor to push their AI budgets.

Irrespective of in the event you’re a skeptic or supporter of AI, this new analysis offers heaps to consider by way of adoption, funding and the way forward for synthetic intelligence as a complete.

Perplexity Continues Its Push To Please Publishers

This month, Perplexity launched a brand new standalone subscription known as Comet Plus. This new tier offers customers entry to premium content material from choose trusted publishers and journalists, with 80% of subscription income paid on to these entities, whereas 20% is reserved for Perplexity’s computing and working prices. Particularly, income is allotted to Perplexity’s companions primarily based on a number of several types of web site visitors, together with:

• Human visits.
• Search citations.
• AI agent actions.

“The income allocation acknowledges the fact that customers now select how they wish to eat info: looking manually, asking for AI-generated solutions, or deploying brokers for advanced duties,” wrote the Perplexity Staff. “Publishers deserve compensation that matches this new actuality.”

The platform is on a mission to create a greater and fairer web, which, in its present state, they are saying is crammed with “low-quality clickbait, high-friction experiences, and limitless blue hyperlinks to human slop and no solutions.”

This looks like an awesome subsequent step and evolution from Preplexity’s earlier Publishers’ Program, which shared income from advert placements that appeared alongside responses citing publishers’ content material.

On the finish of their announcement, they allude that Comet will ultimately turn out to be free for all Perplexity customers. In the event you can’t abdomen one other subscription, you’ll nonetheless get to expertise Comet in due time.

Is the AI Bubble About To Burst? Many Suppose So Following GPT-5 Launch

Dialogue a few bursting AI bubble has been brewing for a very long time. However, ever since OpenAI unveiled GPT-5 on August seventh, 2025, that simmering pot appears to be getting nearer to boiling over.

The brand new mannequin for ChatGPT was met with combined impressions, with many customers saying it carried out worse than earlier variations throughout a number of of its capabilities. This flop adopted early guarantees about GPT-5’s alleged era-defining upgrades, that are kind of seen as simply one other title change by now. However it triggered an uproar of voices concerning the AI bubble, with media shops and thought leaders from throughout discussing their takes on when it can finally burst. Meta — notorious for its AI spending this 12 months — is even purportedly slowing down its actions because it restrategizes for the nth time in 2025.

Whether or not or not we’re shut to a different dot-com-type detonation with AI is unclear. Actionable recommendation for now? Proceed doing what works greatest for your corporation, keep apprised of updates and take motion accordingly.

Have We Lastly Caught As much as Innovation?

As all the time, it’s not possible to know the place all of that is headed for sure. However August occasions have negatively tinged AI discuss maybe greater than ever earlier than. Extra individuals are skeptical, huge manufacturers are allegedly slowing down their AI efforts and, to high it off, MIT says enterprise AI funding seldom brings measurable returns.

The dash towards synthetic basic intelligence (AGI) is inevitable, irrespective of how we really feel about it. However maybe August rekindles an age-old reminder that taking it “sluggish and regular” is nearly all the time higher than speeding, solely to burn out and are available crashing down.

Observe: This text was initially revealed on contentmarketing.ai.